scholarly journals Turbulence Models for Single Phase Flow Simulation of Cyclonic Flotation Columns

Minerals ◽  
2019 ◽  
Vol 9 (8) ◽  
pp. 464 ◽  
Author(s):  
Shiqi Meng ◽  
Xiaoheng Li ◽  
Xiaokang Yan ◽  
Lijun Wang ◽  
Haijun Zhang ◽  
...  

Cyclonic fields are important for cyclonic static microbubble flotation columns (FCSMCs), one of the most important developments in column flotation technology, particularly for separation of fine particles, where the internal flow field has enormous influence on flotation performance. PIV (particle image velocimetry) and CFD (computational fluid dynamics) are the most effective methods to study flow fields. However, data is insufficient for FCSMC flow fields and similar cyclonic equipment, with turbulence model simulations producing different views to measured data. This paper employs an endoscope and PIV to measure axial and cross sections for single-phase swirling flow fields in FCSMCs. We then compare various turbulence model simulations (Reynolds stress model (RSM), standard k-ε, realizable k-ε, and RNG (renormalization group) k-ε) to the measured data. The RSM (Reynolds stress model) predicts cyclonic flow field best in flotation columns with 16.22% average relative velocity deviation. Although the realizable k-ε model has less than 30% relative deviation in radial and tangential directions, axial deviations reach 78.11%. Standard k-ε and RNG k-ε models exhibited approximately 40% and 30% radial and tangential deviation, respectively, and cannot be used even for trend predictions for axial velocity. k-ε models are based on isotropic assumptions with semi-empirical formulas summarized from experiments, whereas RSM fundamentally considers laminar flow and Reynolds stress, and hence is more suitable for anisotropic performance. This study will contribute to flotation column and other cyclonic flow field equipment research.

Author(s):  
Jürgen R. Lücke ◽  
Heinz E. Gallus

The flow field inside an annular compressor cascade is numerically investigated. The mean flow features are complex three-dimensional zones of turbulent separation at hub and shroud at high inflow angles. The flow field is investigated with an implicit three-dimensional Navier-Stokes code. To predict turbulent effects the flow solver includes two different variants of a Low-Re-number k-ϵ-model and an algebraic Reynolds-stress-model. Using the Low-Re-number model the structure of the regions of separated flow are fairly well predicted. However, intensity and size of these zones are too small compared with the experimental data. Better results are produced using the anisotropic algebraic Reynolds-stress-model combined with a stagnation point modification of the turbulent production term. Stucture and intensity of the vortex systems are simulated in more detail. Static pressure distributions and loss contours are in a very good agreement with the experiments.


1999 ◽  
Vol 121 (4) ◽  
pp. 255-262 ◽  
Author(s):  
G. K. Morris ◽  
S. V. Garimella ◽  
J. A. Fitzgerald

The flow field of a normally impinging, axisymmetric, confined and submerged liquid jet is predicted using the Reynolds Stress Model in the commercial finite-volume code FLUENT. The results are compared with experimental measurements and flow visualizations and are used to describe the position of the recirculating toroid in the outflow region which is characteristic of the confined flow field. Changes in the features of the recirculation pattern due to changes in Reynolds number, nozzle diameter, and nozzle-to-target plate spacing are documented. Results are presented for nozzle diameters of 3.18 and 6.35 mm, at jet Reynolds numbers in the range of 2000 to 23,000, and nozzle-to-target plate spacings of 2, 3, and 4 jet diameters. Up to three interacting vortical structures are predicted in the confinement region at the smaller Reynolds numbers. The center of the primary recirculation pattern moves away from the centerline of the jet with an increase in Reynolds number, nozzle diameter, and nozzle-to-target plate spacing. The computed flow patterns were found to be in very good qualitative agreement with experiments. The radial location of the center of the primary toroid was predicted to within ±40 percent and ±3 percent of the experimental position for Re = 2000–4000 and Re = 8500–23000, respectively. The magnitude of the centerline velocity of the jet after the nozzle exit was computed with an average error of 6 percent. Reasons for the differences between the numerical predictions at Re = 2000–4000 and experiments are discussed. Predictions of the flow field using the standard high-Reynolds number k-ε and renormalization group theory (RNG) k-ε models are shown to be inferior to Reynolds stress model predictions.


AIAA Journal ◽  
1997 ◽  
Vol 35 ◽  
pp. 91-98
Author(s):  
Jiang Luo ◽  
Budugur Lakshminarayana

Author(s):  
David Roos Launchbury ◽  
Luca Mangani ◽  
Ernesto Casartelli ◽  
Francesco Del Citto

Abstract In the industrial simulation of flow phenomena, turbulence modeling is of prime importance. Due to their low computational cost, Reynolds-averaged methods (RANS) are predominantly used for this purpose. However, eddy viscosity RANS models are often unable to adequately capture important flow physics, specifically when strongly anisotropic turbulence and vortex structures are present. In such cases the more costly 7-equation Reynolds stress models often lead to significantly better results. Unfortunately, these models are not widely used in the industry. The reason for this is not mainly the increased computational cost, but the stability and convergence issues such models usually exhibit. In this paper we present a robust implementation of a Reynolds stress model that is solved in a coupled manner, increasing stability and convergence speed significantly compared to segregated implementations. In addition, the decoupling of the velocity and Reynolds stress fields is addressed for the coupled equation formulation. A special wall function is presented that conserves the anisotropic properties of the model near the walls on coarser meshes. The presented Reynolds stress model is validated on a series of semi-academic test cases and then applied to two industrially relevant situations, namely the tip vortex of a NACA0012 profile and the Aachen Radiver radial compressor case.


Sign in / Sign up

Export Citation Format

Share Document