Session A36: Turbulence: Flow Generation and Flow Measurements Developments

Turbulence decay in a von Kármán swirling flow

In this study, we investigate the decay of high Reynolds number turbulence at the center of a large von Kármán swirling flow facility (2m height and 2m across). The facility consists of a dodecagonal with two counter-rotating impellers that produces stationary homogenous turbulence near the center of the tank with low mean velocity and intense fluctuations. To measure the decay rate, we run the impellers until the flow reaches a stationary state and then stop them and immediately start measuring the velocity field in a plane at the center of the tank using Stereoscopic Particle Image Velocimetry (S-PIV).  We run the experiment for 3 different Reynolds numbers Re_R=R²ω/ν=2.54×10⁵,1.91×10⁵,1.27×10⁵ where R is the impeller radius, ω is the angular velocity of the impeller and ν is the kinematic viscosity of the flow. Over 30 runs for each Re_R were obtained to ensure reasonable convergence of the turbulent statistics. We calculate the spatially averaged kinetic energy over the field of view (FoV) from the PIV measurements and show that, after a proper normalization, the kinetic energy remains approximately constant for a short period after the impellers are stopped before transitioning to the expected power law decay. Our results consistently follow the power law k∼tⁿ with a decay exponent n≅-1.6 .

    

Exploration of geometric constraints in laboratory-generated homogeneous isotropic turbulence via random jet arrays

Laboratory generation of mean shear free homogeneous isotropic turbulence (HIT) has progressed significantly since the 1950s. Since the creation of the random jet array by Variano et al. (2004), in which grids of randomly-actuated recirculating pumps produce HIT, refined control of turbulence statistics is possible by changing both the geometry of the facility and actuation of the stochastic forcing algorithm. While prior studies have shown turbulence metrics (e.g. turbulent kinetic energy, dissipation, integral length scale) correlate to the mean on-time of the jets specified in the algorithm, questions remain regarding the importance of jet spacing and jet diameter in producing mean shear free HIT of targeted turbulence levels. We perform a laboratory study to explore the relative importance of the arrangement of jets (jet spacing), algorithm parameters (mean on-time; percentage of active jets), and jet outlet velocity on the energetics of the turbulence produced, as well as implications for determining the distance from the jet array at which HIT is generated. We use particle image velocimetry to quantify the flow dynamics produced by 8x8, 12x12, and 16x16 arrays, and we perform extensive dimensional analysis to thoroughly characterize turbulence generation.   

How to generate turbulence with highest Reynolds numbers in the wind tunnel

In order to study objects like buildings, vehicles or wind turbines under the influence of wind fluctuations, the generation of laboratory flows that resemble atmospheric turbulence is of prime importance. This is where active grids come into play, allowing to excite the wind tunnel flow in a user-defined way. With a blockage induced flow design, it is possible to recreate atmospheric flows through their time series or to create single coherent structures such as gusts defined by industrial standards. In addition, it is possible to generate turbulence with large integral length scales through a random driving that follows a stochastic process. Velocity fluctuations with correlation lengths and thus integral scales much larger than the transverse dimensions of the wind tunnel can be generated. By combining active grid excitation with fan speed modulation, it is additionally possible to generate a flow characterized by an inertial range of four decades and an integral Reynolds number of 2×10⁷. By a newly developed active grid it is furthermore possible to vary the turbulent properties over height to mimic height dependencies found in the atmosphere and also to generate a turbulent non-turbulent interface.   

Exploring fields of turbulence generated by an active grid

With the invention of active grids, a tool for the manipulation of flows was developed that has fascinated the experimental turbulence community for decades. Different philosophies for the excitation of active grids have led to a large variety of parametric studies. Often, these experimental studies are carried out using hot-wires to investigate the evolution of small-scale turbulence structures. In contrast, here, we present a study of eight excitation protocols of an active grid with a mesh width of M=10cm that is installed in the wind tunnel at NTNU with dimensions of 1.8m x 2.7m x 10m. Measurements were performed 60M downstream of the active grid at 10m/s using particle image velocimetry with a sampling frequency of 1.5Hz and a field of view of 10M in streamwise and 3M in wall-normal direction. This allows for a detailed exploration of the instantaneous behavior of the velocity field. This will be complemented with analyses of the means of streamwise and wall-normal velocity, turbulence intensity and integral length scales with the objective of quantifying the turbulence evolution based on the excitation of the flow, possibly checking for anomalies in the behavior.   

Understanding the influence of atmospheric flow on scalar-mediated insect orientation behavior

Under natural conditions, many flying insects locate resources vital to survival and reproduction by responding to, and tracking, plumes of scalar cues present in their surroundings (e.g. odors, humidity, temperature, carbon dioxide). The spatio-temporal distribution of the scalar field is a function of turbulent fluctuations, or other unsteady flow phenomena, that introduce a large range of length and time scales into the flow present in the atmospheric boundary layer (ABL) or in-built environments. Yet, we know very little about the behavioral strategies by which insects accomplish this remarkable feat under the turbulent conditions characteristic of natural habitats. Much of our current understanding of the insect behavioral mechanisms come from no-flow, limited flow, or laminar-flow wind tunnel experiments. Addressing this limitation would have several important implications, including deployment of optimized distraction strategies and innovative tracking strategies for autonomous vehicles in hazardous environments. Here, we present the preliminary field measurements of the environmental conditions correlated with mosquito activity, made using very-large scale bubble velocimetry developed in-house. A novel, cost-effective turbulence generating grid has been developed to reproduce the wide range of flow scales to enable controlled wind tunnel studies.   

Homogeneous turbulence in an octagonal tank stirred by four synthetic jet arrays

We inspect high Reynolds number homogeneous turbulence with a negligible mean flow in the octagonal tank created by four synthetic jet arrays. Each array has 40 pumps, and more pumps are placed in the vertical direction to generate a vertically large turbulence region. The individual pump firing zero-net- mass flux jet flow turns on and off repeatedly. The durations of the ON / OFF period are chosen from two different Gaussian distributions. In order to improve homogeneity and isotropy, we investigate the effects of placing a grid and a cylindrical obstacle at each jet exit, whose effects on the jet flow is first characterized by time-resolved particle image velocimetry (PIV) measurements in a single jet flow. The obstacle is designed to increase stirring in the vertical direction while diluting the signature of the jet flow. The turbulent quantities such as turbulent kinetic energy, correlation, structure function, and energy spectra from two experimental configurations, with/without the obstacles, are compared using statistically independent velocity snapshots taken with PIV. We also perform time resolved velocity measurements to understand large scale spatiotemporal velocity correlations within the flow.   

Experimental Characterization of Near Wall Structure in Rotating Turbulence

Turbulent flows in rotating systems play an important role in a variety of fields, including climate systems, ocean currents, etc. Compared to the geophysical applications, which are widely studied by most previous research, the near-wall structures and the effects of rotating-related forces (i.e., Coriolis force and centrifugal force) on boundary layers are unidentified. In this study, a new rotating platform is fabricated to achieve controllable rotating speed up to 55 rpm. The vibration of the rotation platform is minimized with machined alignment parts and a turntable bearing to avoid effects on turbulence. The flow features at different rotation numbers are investigated. The 3D flow structures are examined with an onboard 4-camera tomographic PIV system. The centrifugal force and Coriolis force introduce asymmetry in the main flow velocity and the Reynolds stress distribution. They also affect the length scale of the boundary layer as well as the eddy structures.   

On the Spectral Response of a Horizontal Axis Tidal Turbine Subjected to Anisotropic Grid Turbulence

The conversion dynamics from turbulent velocity fluctuations to turbine output are not well understood, particularly for tidal turbines operating in much more severe inflow conditions than their counterparts, like wind energy. These fluctuations have been shown to cause intermittency in the electrical grid for a wind farm creating a need for control devices. To address some of those deficiencies, we have developed a Tidal Turbulence Test Facility (T³F) at Lehigh University which can mimic several turbulence parameters at high-energy tidal sites using a Makita-type active grid. We will present results from our experimental campaign where we investigate the spectral behavior of power and thrust output of a 1:20 model scale tidal turbine subjected to anisotropic inflow conditions. We perform simultaneous upstream velocity (1D from rotor plane) and turbine measurement for different tip-speed ratios. Four inflow conditions are tested: quasi-laminar (QL), double random (DR1) with elevated turbulence intensity (Ti), synchronous (SY1) with Ti similar to DR1 but dominant scales at the grid operating frequency (1Hz), and shear inflow (SH) mimicking the shear profile of tidal sites. The detailed results will be presented and compared based on the power spectral density for each case.

    

Can a turbulence experiment be repeated ?

Using time-reloved PIV, we make movies of a 2D slice of the velocity field behind an impulsively started submerged plate in a water basin. The experiment is exactly repeated 42 times using a robot. The flow is characterized by vortices. The two counter-rotating vortices that are shed initially merge and become turbulent. The outcome of repeated experiments diverges as time progresses. This divergence can be predicted from a single experiment by computing its finite-time Lyapunov field. Remarkably, the initial symmetry of the opposing large-scale circulation remains conserved over the entire trajectory of the plate. Yes, a turbulence experiment can be repeated, but not its outcome. However, in this experiment, the discrepancy between results takes on a remarkable shape.