Numerical Simulation of the Flow in the NASA Low-Speed Centrifugal Compressor

The performance of different low-Reynolds number turbulence models applied to the simulation of a flow in the NASA low-speed compressor is described. The flow solver utilizes a structured multiblock grid with Cartesian velocity components in a rotating coordinate system. The inviscid fluxes are calculated using Roe ' s scheme. An implicit solution method is applied and a multigrid cycling is used in order to accelerate the convergence. In this study the Baldwin-Lomax model, Chien 's k ε model and a full Reynolds-stress closure by Speziale et al. are applied. A detailed comparison between the experimental and computational velocity fields and pressure distributions is made. N o m e n c l a t u r e A Jacobian matrix of the flux vector CFL Courant number D diffusion of the Reynolds stresses Ε total energy per unit volume F,G,H flux vectors in the x,y and z-directions G production of the kinetic energy of turbulence due to rotation Q source term vector Ρ production of the kinetic energy of turbulence Τ temperature Tu turbulence intensity, Tu = j-k/ub U vector of conservative variables Uh bulk velocity (= 36.1 m/s) W relative velocity; vorticity rate tensor ij,k grid coordinate directions k kinetic energy of turbulence, m/ms nondimensional shroud meridional distance ρ static pressure pxli, pressure at standard-day sea-level conditions (=101325 N/m) r location vector t time Uj velocity components in the /-direction u, v, w velocity components in the x, y and zdirections x, y, ζ Cartesian coordinates y + non-dimensional normal distance from the surface Φ pressure-strain correlation Ω angular velocity Ω angular velocity rate vector γ ratio of specific heats; constant in pressurestrain correlation bij Kronecker's delta ε dissipation of the kinetic energy of turbulence ε,β permutation tensor η efficiency μ dynamic viscosity ρ density σ Schmidt's number φ scalar variable; part of the pressure-strain correlation Subscripts TS total to static Τ turbulent conditions * Research Scientist, tel: +358 (9) 451 3949, e-mail: Patrik.Rautaheimo@hut.fi " Research Engineer, tel: +358 (9)451 3425, e-mail: Esa.Salminen@hut.fi "* Professor, tel: +358 (9) 451 3426, e-mail: Timo.Siikonen@hut.fl

[1]  H. Lomax,et al.  Thin-layer approximation and algebraic model for separated turbulent flows , 1978 .

[2]  T. Gatski,et al.  Modelling the pressure–strain correlation of turbulence: an invariant dynamical systems approach , 1991, Journal of Fluid Mechanics.

[3]  K. Chien,et al.  Predictions of Channel and Boundary-Layer Flows with a Low-Reynolds-Number Turbulence Model , 1982 .

[4]  D. Vandromme Turbulence modeling for compressible flows and implementation in Navier-Stokes solvers , 1991 .

[5]  Michael D. Hathaway,et al.  Laser Anemometer Measurements of the Three-Dimensional Rotor Flow Field in the NASA Low-Speed Centrifugal Compressor , 1995 .

[6]  Timo Siikonen,et al.  An application of Roe's flux‐difference splitting for k‐ϵ turbulence model , 1995 .

[7]  Arthur Rizzi,et al.  Navier-Stokes stall predictions using an algebraic Reynolds-stress model , 1992 .

[8]  E. Venkatapathy,et al.  Multi-dimensional formulation of CSCM - An upwind flux difference eigenvector split method for the compressible Navier-Stokes equations , 1983 .

[9]  Nobuyuki Shima,et al.  A Reynolds-stress model for near-wall and low-Reynolds-number regions , 1988 .

[10]  D. Holmes,et al.  A Three-Dimensional Euler Solver for Turbomachinery Blade Rows , 1985 .

[11]  Budugur Lakshminarayana,et al.  Three-Dimensional Navier–Stokes Computation of Turbine Nozzle Flow With Advanced Turbulence Models , 1997 .

[12]  P. Roe Approximate Riemann Solvers, Parameter Vectors, and Difference Schemes , 1997 .

[13]  G. D. van Albada,et al.  A comparative study of computational methods in cosmic gas dynamics , 1982 .

[14]  Antony Jameson,et al.  Multigrid solution of the Euler equations using implicit schemes , 1985 .