Assessment of Turbulence Modeling for Compressible Flow Around Stationary and Oscillating Cylinders


Book Description

A numerical analysis of the flow over stationary and transversely oscillating circular cylinders at Reynolds numbers of 3900 and 3600, respectively, is undertaken to assess various turbulence modeling techniques for the simulation of vortex shedding phe-nomena. Four turbulence models are considered; namely the one-equation Spalart-Allmaras model [Spalart & Allmaras (1994)] and the k-T model by Speziale et al. (1992) for URANS closure, the constant-coefficient Smagorinsky-Lilly- subgrid-scale model for Large Eddy Simulations, and the adaptive k-T, model proposed by Magag?nato Gabi (2002) for Very Large Eddy Simulations. A key contribution of this work is comparison of results obtained with the same numerical procedure, discretization algorithms, and artificial dissipation but different turbulence modeling techniques in order to properly differentiate between errors due to numerical and to modeling aspects. Through the study of the flow around the stationary cylinder, it is shown that the use of an inadequate time step has a small effect on global average quantities, but a noticeable impact on the pressure coefficient around the cylinder, as well as on the evolution of velocity along the centerline, thus showing that simulations with too large a time step are unable to properly resolve the recirculation zone and wake. Global average quantities are found to be relatively insensitive to three-dimensional resolution. Detailed analysis of three-dimensional URANS simulations reveals that the Spalart-Allmaras model is unable to properly predict the location of the separation point. separation being delayed significantly. This results in too small a mean recircula?tion zone, under-estimated back-pressure. and up to 25% over-estimation of the drag. Yet, this simple model provides an accurate value for the Strouhal number and good fluctuating velocity profiles. The k-T Speziale turbulence model predicts all global quantities accurately, and yields good velocity profiles along the wake as well as an adequate pressure distribution on the cylinder wall. Large and Very Large Eddy Simulations of the flow around a stationary cylinder reveal an important three-dimensionality, and the formation on the upper and lower surfaces of two secondary eddies in addition to the two large vortices. Furthermore, the LES properly captures the dynamics in the laminar boundary layer as reflected by the skin friction values even though it makes use of a constant coefficient Smagorinsk-y?subgrid-scale model. In the study of the transversely oscillating cylinder with two-dimensional URANS k-T Speziale simulations, the lock-in region starts at significantly lower motion fre-quencies than observed in experimental results, while the phase shift was not observed. No numerical studies close to this Reynolds number are available in the literature. and further investigation is required.




Assessment of Turbulence Modeling for Compressible Flow Around Stationary and Oscillating Cylinders


Book Description

A numerical analysis of the flow over stationary and transversely oscillating circular cylinders at Reynolds numbers of 3900 and 3600, respectively, is undertaken to assess various turbulence modeling techniques for the simulation of vortex shedding phe-nomena. Four turbulence models are considered; namely the one-equation Spalart-Allmaras model [Spalart & Allmaras (1994)] and the k-T model by Speziale et al. (1992) for URANS closure, the constant-coefficient Smagorinsky-Lilly- subgrid-scale model for Large Eddy Simulations, and the adaptive k-T, model proposed by Magag?nato Gabi (2002) for Very Large Eddy Simulations. A key contribution of this work is comparison of results obtained with the same numerical procedure, discretization algorithms, and artificial dissipation but different turbulence modeling techniques in order to properly differentiate between errors due to numerical and to modeling aspects. Through the study of the flow around the stationary cylinder, it is shown that the use of an inadequate time step has a small effect on global average quantities, but a noticeable impact on the pressure coefficient around the cylinder, as well as on the evolution of velocity along the centerline, thus showing that simulations with too large a time step are unable to properly resolve the recirculation zone and wake. Global average quantities are found to be relatively insensitive to three-dimensional resolution. Detailed analysis of three-dimensional URANS simulations reveals that the Spalart-Allmaras model is unable to properly predict the location of the separation point. separation being delayed significantly. This results in too small a mean recircula?tion zone, under-estimated back-pressure. and up to 25% over-estimation of the drag. Yet, this simple model provides an accurate value for the Strouhal number and good fluctuating velocity profiles. The k-T Speziale turbulence model predicts all global quantities accurately, and yields good velocity profiles along the wake as well as an adequate pressure distribution on the cylinder wall. Large and Very Large Eddy Simulations of the flow around a stationary cylinder reveal an important three-dimensionality, and the formation on the upper and lower surfaces of two secondary eddies in addition to the two large vortices. Furthermore, the LES properly captures the dynamics in the laminar boundary layer as reflected by the skin friction values even though it makes use of a constant coefficient Smagorinsk-y?subgrid-scale model. In the study of the transversely oscillating cylinder with two-dimensional URANS k-T Speziale simulations, the lock-in region starts at significantly lower motion fre-quencies than observed in experimental results, while the phase shift was not observed. No numerical studies close to this Reynolds number are available in the literature. and further investigation is required.




Assessment of Turbulence Modeling for Compressible Flow Around Stationary and Oscillating Cylinders


Book Description

A numerical analysis of the flow over stationary and transversely oscillating circular cylinders at Reynolds numbers of 3900 and 3600, respectively, is undertaken to assess various turbulence modeling techniques for the simulation of vortex shedding phe-nomena. Four turbulence models are considered; namely the one-equation Spalart-Allmaras model [Spalart & Allmaras (1994)] and the k-T model by Speziale et al. (1992) for URANS closure, the constant-coefficient Smagorinsky-Lilly- subgrid-scale model for Large Eddy Simulations, and the adaptive k-T, model proposed by Magagnato Gabi (2002) for Very Large Eddy Simulations. A key contribution of this work is comparison of results obtained with the same numerical procedure, discretization algorithms, and artificial dissipation but different turbulence modeling techniques in order to properly differentiate between errors due to numerical and to modeling aspects. Through the study of the flow around the stationary cylinder, it is shown that the use of an inadequate time step has a small effect on global average quantities, but a noticeable impact on the pressure coefficient around the cylinder, as well as on the evolution of velocity along the centerline, thus showing that simulations with too large a time step are unable to properly resolve the recirculation zone and wake. Global average quantities are found to be relatively insensitive to three-dimensional resolution. Detailed analysis of three-dimensional URANS simulations reveals that the Spalart-Allmaras model is unable to properly predict the location of the separation point. separation being delayed significantly. This results in too small a mean recirculation zone, under-estimated back-pressure. and up to 25% over-estimation of the drag. Yet, this simple model provides an accurate value for the Strouhal number and good fluctuating velocity profiles. The k-T Speziale turbulence model predicts all global quantities accurately, and yields good velocity profiles along th.







Compressibility, Turbulence and High Speed Flow


Book Description

This book introduces the reader to the field of compressible turbulence and compressible turbulent flows across a broad speed range through a unique complimentary treatment of both the theoretical foundations and the measurement and analysis tools currently used. For the computation of turbulent compressible flows, current methods of averaging and filtering are presented so that the reader is exposed to a consistent development of applicable equation sets for both the mean or resolved fields as well as the transport equations for the turbulent stress field. For the measurement of turbulent compressible flows, current techniques ranging from hot-wire anemometry to PIV are evaluated and limitations assessed. Characterizing dynamic features of free shear flows, including jets, mixing layers and wakes, and wall-bounded flows, including shock-turbulence and shock boundary-layer interactions, obtained from computations, experiments and simulations are discussed. Describes prediction methodologies including the Reynolds-averaged Navier Stokes (RANS) method, scale filtered methods and direct numerical simulation (DNS) Presents current measurement and data analysis techniques Discusses the linkage between experimental and computational results necessary for validation of numerical predictions Meshes the varied results of computational and experimental studies in both free and wall-bounded flows to provide an overall current view of the field




Turbulence Model Assessment for High Speed Compressible Flows


Book Description

Modeling turbulence is an important aspect of high speed flows from a solution accuracy point of view. This present research work focuses on the assessment of different turbulence models in predicting shock-wave turbulent boundary layer interactions induced by a compression ramp. This case is unique due to the fact that unlike earlier works on this topic, it focuses on non-separating boundary layer. Five turbulence models were assessed and are compared according to the prediction of pressure profiles, skin friction coefficient profiles and boundary layer velocity profiles.




Turbulence Modelling Approaches


Book Description

Accurate prediction of turbulent flows remains a challenging task despite considerable work in this area and the acceptance of CFD as a design tool. The quality of the CFD calculations of the flows in engineering applications strongly depends on the proper prediction of turbulence phenomena. Investigations of flow instability, heat transfer, skin friction, secondary flows, flow separation, and reattachment effects demand a reliable modelling and simulation of the turbulence, reliable methods, accurate programming, and robust working practices. The current scientific status of simulation of turbulent flows as well as some advances in computational techniques and practical applications of turbulence research is reviewed and considered in the book.







Advanced Approaches in Turbulence


Book Description

Front Cover -- Advanced Approaches in Turbulence -- Copyright -- Contents -- Contributors -- Preface -- 1 Basics of turbulence -- 1.1 Introduction -- 1.2 Eddy diffusion -- 1.3 Scales of turbulence -- 1.3.1 Isotropic decay -- 1.3.2 Stretching and diffusion of vorticity -- 1.4 Spectral equations -- 1.4.1 Isotropic turbulence -- 1.4.2 Shear and streaks -- 1.5 Averaged equations -- 1.5.1 Jets -- 1.5.2 Boundary layer -- 1.6 The form of turbulence models -- 1.6.1 Two equation models -- 1.6.2 Reynolds stress transport -- 1.7 Conclusion -- References -- 2 Direct numerical and large-eddy simulation of complex turbulent flows -- 2.1 Introduction -- 2.2 Error as a function of scale -- 2.2.1 Modified wavenumber -- 2.2.2 Nonlinear sources of error -- 2.2.3 Time advancement error as a function of scale -- 2.3 Analysis of numerical errors in large-eddy simulation using statistical closure theory -- 2.3.1 EDQNM closure -- 2.3.2 EDQNM-LES and the inclusion of numerical error -- 2.3.3 EDQNM model -- 2.3.4 Relative magnitudes of error -- 2.4 Simulations in complex geometries -- 2.4.1 Decay of isotropic turbulence -- 2.4.2 Gas turbine combustor -- 2.5 Simulating the flow around moving bodies -- 2.5.1 Fluid phase -- 2.5.2 Solid phase -- 2.5.3 The effects of interpolation -- 2.5.4 Particles in a turbulent channel -- 2.6 What is a 'canonical' flow? -- 2.6.1 Jets in crossflow -- 2.6.2 DNS of turbulent channel flow over random rough surfaces -- 2.7 The analysis of 'big data' -- 2.7.1 DMD of large datasets and numerical error -- 2.7.2 Analysis of wall-pressure fluctuation sources in turbulent channel flow -- 2.8 Bridging the Reynolds number divide -- 2.9 Concluding remarks -- Acknowledgments -- References -- 3 Large-eddy simulations -- 3.1 Introduction -- 3.1.1 Motivation -- 3.2 Governing equations -- 3.2.1 Filtering.