0

### Research Papers: Flows in Complex Systems

J. Fluids Eng. 2017;139(5):051101-051101-12. doi:10.1115/1.4035804.

The influence of ground effect on the wake of a high-speed train (HST) is investigated by an improved delayed detached-eddy simulation. Aerodynamic forces, the time-averaged and instantaneous flow structure of the wake are explored for both the stationary ground and the moving ground. It shows that the lift force of the trailing car is overestimated, and the fluctuation of the lift and side force is much greater under the stationary ground, especially for the side force. The coexistence of multiscale vortex structures can be observed in the wake along with vortex stretching and pairing. Furthermore, the out-of-phase vortex shedding and oscillation of the longitudinal vortex pair in the wake are identified for both ground configurations. However, the dominant Strouhal number of the vortex shedding for the stationary and moving ground is 0.196 and 0.111, respectively, due to the different vorticity accumulation beneath the train. A conceptual model is proposed to interpret the mechanism of the interaction between the longitudinal vortex pair and the ground. Under the stationary ground, the vortex pair embedded in a turbulent boundary layer causes more rapid diffusion of the vorticity, leading to more intensive oscillation of the longitudinal vortex pair.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051102-051102-11. doi:10.1115/1.4035640.

Renewable energy sources (RES) have reached 23.7% of the worldwide electrical generation production in 2015. The hydraulic energy contribution amounts to 16.6% and comes mainly form large-scale hydropower plants, where Francis turbines represents 60% of the generating units. However, the future massive development of RES will require more advanced grid regulation strategies that may be achieved by increasing the operation flexibility of the Francis generating units. Part load operating condition of these turbines is hindered by pressure fluctuations in the draft tube of the machine. A precessing helical vortex rope develops in this condition, which imperils the mechanical structure and limits the operation flexibility of these turbines. A thorough description of the physical mechanism leading to the vortex rope is a prerequisite to develop relevant flow control strategies. This work, based on a linear global stability analysis of the time-averaged flow field, including a turbulent eddy viscosity, interprets the vortex rope as a global unstable eigenmode. In close resemblance to spiral vortex breakdown, a single-helix disturbance develops around the time-averaged flow field and growths in time to finally form the vortex rope. The frequency and the structure of this unstable linear disturbance are found in good agreement with respect to the three-dimensional (3D) numerical flow simulations.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051103-051103-14. doi:10.1115/1.4035638.

Severe vibrations induced by flow instabilities in the nuclear reactor coolant pump (RCP) are detrimental to the safe operation of the pump. Due to the particular spherical casing in the RCP, the internal flow structures are extremely ambiguity and complicated. The goal of the present work is to shed comprehensive light on the unsteady flow structures and its correlation with the pressure pulsations by using large eddy simulation (LES) method of the RCP. The vorticity distribution and the shedding vortex from the blade trailing edge are depicted in detail. Furthermore, the internal correlations between the flow unsteadiness and pressure pulsation are illustrated in some special regions of the RCP. Evidently, some main excitation components in the pressure spectra are excited by the shedding vortex. Besides, components at blade passing frequency (fBPF) are closely associated with rotor–stator interaction between the wake flow from the impeller outlet and unsteadiness vortexes shedding from the diffuser blade trailing edge. It is thought to be that the pressure pulsations of the RCP are closely associated with the corresponding vorticity distribution and the unsteady vortex shedding effect.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051104-051104-13. doi:10.1115/1.4035462.

The flow field in a linear cascade of highly loaded turbine nozzle guide vanes (NGVs) has been numerically investigated at low and high-subsonic regime, i.e., exit isentropic Mach number of M2is = 0.2 and 0.6, respectively. Extensive experimental data are available for an accurate assessment of the numerical procedure. Aerodynamic measurements include not only vane loading and pressure drop in the wake but also local flow features such as boundary layer behavior along both pressure and suction sides of the vane, as well as secondary flow structures downstream of the trailing edge (TE). Simulations were performed by using two computational fluid dynamics (CFD) codes, a commercial one and an open-source based in-house code. Besides computations with the well-established shear-stress transport (SST) k–ω turbulence model assuming fully turbulent flow, transition models were taken into account in the present study. The original version of the γ–Reθ model of Menter was employed. Suluksna–Juntasaro correlations for transition length (Flenght) and transition onset (Fonset) were also tested. The main goal was to establish essential ingredients for reasonable computational predictions of the cascade aerodynamic behavior, under both incompressible and compressible regime. This study showed that transition modeling should be coupled with accurate profiles of inlet velocity and turbulence intensity to get a chance to properly quantify aerodynamic losses via CFD method. However, additional weaknesses of the transition modeling have been put forward when increasing the outlet Mach number.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051105-051105-9. doi:10.1115/1.4035634.

The dispensing behavior of a piezo-actuated micro-valve that closes the gap between the nanoliter range (e.g., inkjet technology) and the microliter range (e.g., standard displacement technology) has been investigated by experimental and numerical means. Water and different Newtonian model fluids with defined fluid properties were utilized for experimental characterization. The dispensed amount per single dispensing event could be freely adjusted from a few nanoliters to several hundred microliters showing the large working range and flexibility of the micro-valve, while maintaining a high accuracy with a low relative standard deviation. A correlation between fluid properties, dispensing parameters, and the resulting steady-state mass flow was established, showing good consistency of the experimental data. Furthermore, a three-dimensional numerical model for the quantitative simulation of the micro-valve's dispensing behavior regarding fluid mass flow was developed and validated, showing a high degree of correspondence between the experiments and simulations. Investigations of the transient behavior after the opening of the micro-valve revealed a nonlinear relationship between the valve opening time and dispensed mass for short opening times. This behavior was dependent on the working pressure but independent of the type of fluid.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051106-051106-8. doi:10.1115/1.4035639.

The effect of apex flap and tip flap, deflected both independently and jointly, on the vortex flow and lift generation of a 65 deg-sweep delta wing was investigated experimentally. The drooped apex flap produced a higher lift at medium-to-high angle of attack regime and also a delayed stall. The anhedral (introduced by the downward tip flap) generally promoted lift increment, whereas dihedral had the opposite effect. Meanwhile, the joint apex and tip flap deflection gave a delayed leading-edge vortex (LEV) breakdown and an enhanced lift. The LEVs were generally drawn closer to the wing upper surface, while being pushed further away from the wing centerline by the application of apex flap and tip flap. The flap also modified the vorticity distribution in the LEV; the bursting behavior was, however, not affected. Dye-injection flow visualization and particle image velocimetry (PIV) measurements of the vortex flow were also discussed.

Commentary by Dr. Valentin Fuster

### Research Papers: Fundamental Issues and Canonical Flows

J. Fluids Eng. 2017;139(5):051201-051201-7. doi:10.1115/1.4035632.

An experimental investigation was carried out to study the turbulent flow behind passive grids in a subsonic wind tunnel. The enhanced level of turbulence was generated by five wicker metal grids with square meshes and different parameters (diameter of the grid rod d = 0.3 to 3 mm and the grid mesh size M = 1 to 30 mm). The velocity of the flow was measured by means of a one-dimensional hot-wire probe. For this purpose, skewness, kurtosis, and transverse variation of the velocity fluctuations were determined, obtaining knowledge of the degree of turbulence isotropy and homogeneity in the flow behind grids of variable geometry, for different incoming velocities U = 4, 6, 10, 15, 20 m/s. Approximately, the isotropic and homogeneous turbulence was obtained for x/M > 30. Next, several correlations for turbulence degeneration law were tested. Finally, as the main goal of the study, impact of turbulence intensity on bypass laminar–turbulent transition parameters (transition inception, shape parameter, and the length of the transition region) on a flat plate was investigated. Parameter ITum was created as an integral taken from the leading edge of the plate to the transition inception, divided by the distance from the leading edge to the transition inception, expressing in this way the averaged value of turbulence intensity.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051202-051202-6. doi:10.1115/1.4035637.

This paper presents an analytical solution of the momentum equation for the unsteady motion of fluids in circular pipes, in which the kinematic viscosity is allowed to change arbitrarily in time. Velocity and flow rate are expressed as a series expansion of Bessel and Kelvin functions of the radial variable, whereas the dependence on time is expressed as Fourierlike series. The analytical solution for the velocity is compared with the direct numerical solution of the momentum equation in a particular case, verifying that the difference between analytical and numerical values of axial velocity is less than 1%, except near the discontinuity of the applied pressure gradient, where the typical behavior due to the Gibbs phenomenon is to be noted.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051203-051203-6. doi:10.1115/1.4035633.

This paper investigates the self-similarity properties in the far downstream of high Reynolds number turbulent wake flows. The growth rate of the wake layer width, $dδ/dx$; the decaying rate of the maximum velocity defect, $dUs/dx$; and the scaling for the maximum mean transverse (across the stream) velocity, Vmax, are derived directly from the self-similarity of the continuity equation and the mean momentum equation. The analytical predictions are validated with the experimental data. Using an approximation function for the mean axial flow, the self-similarity analysis yields approximate solutions for the mean transverse velocity, V, and the Reynolds shear stress, $T=−〈uv〉$. Close relations among the shapes of U, V, and T are revealed.

Commentary by Dr. Valentin Fuster

### Research Papers: Multiphase Flows

J. Fluids Eng. 2017;139(5):051301-051301-12. doi:10.1115/1.4035300.

Computational analyses of fluid flow through packed pebble bed domains using the Reynolds-averaged Navier–Stokes (RANS) framework have had limited success in the past. Because of a lack of high-fidelity experimental or computational data, optimization of Reynolds-averaged closure models for these geometries has not been extensively developed. In the present study, direct numerical simulation (DNS) was employed to develop a high-fidelity database that can be used for optimizing Reynolds-averaged closure models for pebble bed flows. A face-centered cubic (FCC) domain with periodic boundaries was used. Flow was simulated at a Reynolds number of 9308 and cross-verified by using available quasi-DNS data. During the simulations, low-frequency instability modes were observed that affected the stationary solution. These instabilities were investigated by using the method of proper orthogonal decomposition, and a correlation was found between the time-dependent asymmetry of the averaged velocity profile data and the behavior of the highest-energy eigenmodes. Finally, the effects of the domain size and the method of averaging were investigated to determine how these parameters influenced the stationary solution. A violation of the ergodicity assumption was observed.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051302-051302-13. doi:10.1115/1.4035635.

Poiseuille flows at two Reynolds numbers (Re) 2.5 × 10−2 and 5.0 are simulated by two different smoothed particle hydrodynamics (SPH) schemes on regular and irregular initial particles' distributions. In the first scheme, the viscous stress is calculated directly by the basic SPH particle approximation, while in the second scheme, the viscous stress is calculated by the combination of SPH particle approximation and finite difference method (FDM). The main aims of this paper are (a) investigating the influences of two different schemes on simulations and reducing the numerical instability in simulating Poiseuille flows discovered by other researchers and (b) investigating whether the similar instability exists in other cases and comparing results with the two viscous stress approximations. For Re = 2.5 × 10−2, the simulation with the first scheme becomes instable after the flow approaches to steady-state. However, this instability could be reduced by the second scheme. For Re = 5.0, no instability for two schemes is found.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051303-051303-13. doi:10.1115/1.4035762.

Laboratory experiments were conducted to study the dynamics of sand jets passing through two immiscible fluids. Different oil layer thicknesses, nozzle diameters, and sand masses were employed. Evolution of oily sand jets with time was investigated using image processing and boundary visualization techniques. Different shapes of the frontal head and trailing wave section were observed and cloud formation was classified into different categories based on Reynolds number, normalized oil layer thickness, and evolution time. It was found that the effect of Reynolds number on evolution of oily sand jets was more significant than the other parameters. Width and frontal velocity of oily sand jets were measured at different times. It was observed that oily sand jets became unstable after a distance of ten times larger than the nozzle diameter. Instability of oily sand jets caused intense spreading with a spreading rate of 0.4. The thin layer of oil encapsulated sand cluster was ruptured due to excess shear stress and caused bursting of the frontal head into a cloud of sand particles. Three different bursting mechanisms were observed and a correlation was found between the densimetric Froude number and the normalized bursting time. Data mining and boundary visualization techniques were used to model oily sand jets. Model trees were developed to classify and predict the growth of oily sand jets at different conditions. Modeling results indicated that the Model tree can predict the growth of sand jets with an uncertainty of ±8.2%, ±6.8%, and ±8.7% for width, velocity, and distance, respectively.

Commentary by Dr. Valentin Fuster

### Research Papers: Techniques and Procedures

J. Fluids Eng. 2017;139(5):051401-051401-10. doi:10.1115/1.4035461.

Wind loads on structures and the wind environment around buildings are based on tests in boundary layer wind tunnels with corresponding scale parameters. The lower part of the troposphere boundary layer was simulated inside a small wind tunnel located at the Wind Engineering Centre of the Université de Moncton. The correct scale ratios of the boundary layer thickness combined with the roughness height are two of the most important scales to match. For small wind tunnels, roughness parameters related to the model boundary layer can be difficult to measure since scale ratios for wind load studies are expected to be in the range of 400–1000. Oil-film interferometry was used to determine the roughness parameters (shear stress, friction velocity, and roughness height) of the forced turbulent boundary layer inside the wind tunnel. In this work, the International Organization for Standardization (ISO) Guide to Expression of Uncertainty in Measurements was used to evaluate the standard uncertainty of the roughness parameters on the bottom wall of the wind tunnel. The standard uncertainty of the roughness parameters depends strongly on the oil viscosity and on the accurate measurement of the fringe spacing. Results show that the standard uncertainty of the shear stress and friction velocity determined by the interferometry technique can be less than 5% when the oil viscosity and the fringe spacing can be accurately measured with a standard uncertainty lower than 4% and 1%, respectively.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 2017;139(5):051402-051402-10. doi:10.1115/1.4035806.

By combining a physical model and sensor outputs in an inverse transport-diffusion-reaction strategy, an accurate concentration cartography may be obtained. The paper addresses the influence of discretization errors, flow uncertainties, and measurement noise on the concentration field reconstruction process. We consider a key element of a drinking water network, i.e., a pipe junction, where Reynolds and Peclet numbers are approximately 2000 and 1000, respectively. We show that a 10% error between the reference concentration field and the reconstructed concentration field may be obtained using a coarse discretization. Nevertheless, to keep the error below 10%, a fine concentration discretization is required. We also detail the influence of the flow approximation on the concentration reconstruction process. The flow modeling error obtained when the exact Navier–Stokes flow is approximated by a Stokes flow may lead to a 40% error in the reconstructed concentration. However, if the flow field is obtained from the full set of Navier–Stokes equations, we show that the error may be less than 5%. Then, we observe that the quality of the reconstructed concentration field obtained with the proposed inverse technique is not deteriorated when sensor outputs have a normal distribution noise variance of few percents. Finally, a good engineering practice would be to stop the reconstruction process according to an extended discrepancy principle including modeling and measurement errors. As shown in the paper, the quality of the reconstructed field declines after reaching the threshold of the modeling error.

Commentary by Dr. Valentin Fuster

### Technical Brief

J. Fluids Eng. 2017;139(5):054501-054501-3. doi:10.1115/1.4035761.

Chauvenet's criterion is commonly used for rejection of outliers from sample datasets in engineering and physical science research. Measurement and uncertainty textbooks provide conflicting information on how the criterion should be applied and generally do not refer to the original work. This study was undertaken to evaluate the efficacy of Chauvenet's criterion for improving the estimate of the standard deviation of a sample, evaluate the various interpretations on how it is to be applied, and evaluate the impact of removing detected outliers. Monte Carlo simulations using normally distributed random numbers were performed with sample sizes of 5–100,000. The results show that discarding outliers based on Chauvenet's criterion is more likely to have a negative effect on estimates of mean and standard deviation than to have a positive effect. At best, the probability of improving the estimates is around 50%, which only occurs for large sample sizes.

Commentary by Dr. Valentin Fuster