摘要:

A flux formulation using a projected 2D Roe Riemann solver on unstructured grids (R2D Solver) is introduced for solving the Navier-Stokes equations and is applied to calculations of axisymmetric laminar near-wake flows behind a spherical-conical body. The numerical framework was first developed by P. L. Roe et al, in the late eighties. They looked for unsteady solutions to Euler's equations using a rather simple but exact three state linearization on triangular grids and decomposing the solution using some effective wave models. Our approach differs from their techniques by constructing a second order accurate and conservative flux functions under the well-known classical finite volume formulation. However, our Riemann Solver is obtained by a suitable linearization procedure upon all three prescribed nodal values given on each triangle. Our numerical method is applied to a Mach 4.3 flow problem for refined unstructured triangular grid behind the body. Numerical results indicate that our technique is stable, accurate and converges successfully to a stationary solution as the cell size is reduced from the coarse lo the finest grid.

ISSN:1061-8562    

DOI:10.1080/10618569408904505

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

A spectral method is developed based on the primitive variables for the time-dependent solution of the flow and the temperature past a spherical droplet. Both Chebyshev and Legendre polynomials are used to expand the velocity, pressure, and temperature in the radial and angular directions, respectively. The fractional time-stepping method suggested by Orszag (Orszag et al., 1980) is used for solving the flow and the pressure fields. Euler backward differencing is used for the integration of the energy equation. The computed steady-state drag coefficients are compared to those found in the literature for Reynolds numbers in the range from 0.5 to 50 for both the continuous and the dispersed phase. The transient drag coefficients and Nusselt numbers are compared with our previous study using a stream function-vorticity formulation (Nguyen et al, 1993). The comparison indicates that the present model is capable of predicting the correct nature of the flow and heat transfer associated with a droplet.

ISSN:1061-8562    

DOI:10.1080/10618569408904506

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Analysis of the mixed convection heat transfer in a uniformly heated horizontal pipe, first considered by Morton[Q. J. Mech. & Appl, Math.12(1959) 410], is extended with the aid of algebraic or symbolic computation packages. A set of four coupled, nonlinear partial differential equations that describe the effect of buoyancy-driven secondary flow on the pressure-driven axial flow are solved using a regular perturbation series expansion. A scale analysis suggests the product of Rayleigh and Reynolds numbers as the natural perturbation parameter and the series is constructed around the basic Poiseuille flow. In addition, the solutions also depend on the Prandtl number. The nonlinear problem is thus converted into an infinite series of linear partial differential equations to be solved sequentially. Using a further coordinate transformation into complex coordinates, solution of the infinite series of linear problems is reduced to one of evaluating quadratures, a task which is ideally suited for symbolic computation packages such as REDUCE, MAPLE, or MATHEMATICA. As an example the listing of a MAPLE program to solve the present problem to any desired order is appended. Although the symbolic computation program developed here is capable of solving the problem up to any desired order, the speed and memory limitations of a SUN workstation allowed us to generate analytical expressions for stream function, axial velocity and temperature only up to 10th order. The results provided by Morton for the first 3 terms are in agreement (except for some minor typographical errors) with our computer generated results. Primarily, we illustrate how the tedium of algebraic calculations can be relegated to algebraic computing machines. With the availability of faster CPU and larger memory machines, such problems can be routinely solved analytically to any desired accuracy.

ISSN:1061-8562    

DOI:10.1080/10618569408904507

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

A numerical approach for calculating self-excited unstable aerodynamic/aeroelastic behaviours in lurbo-machinery blading systems is described. The flow field through a cascade section with multiple blade passages (whole annulus) is governed by the 2-D unsteady Navier-Stokes equations, which are spatially discretized by using a finite volume scheme. The motion of each blade section is governed by simple rigid body spring-mass structural dynamic equations with two degrees of freedom (“bending” and “torsion”). The coupled aeroelastic equations are simultaneously integrated in time by using the 4-stage Runge-Kutta scheme. Examples of calculated results for blade rows with rotating stall and stall Mutter are presented.

ISSN:1061-8562    

DOI:10.1080/10618569408904508

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

A numerical scheme for solving the shallow-water equations is presented. An analogy is made between flows governed by shallow-water equations and the Euler system of equations used in gas dynamics. An emphasis is placed on the difference presented by the bathymetry in hydraulic systems. The discretization of the governing equations is based on Roe's flux difference-splitting solver, initially developed for solving inviscid compressible flows. The spatial discretization is handled within a finite-volume context by using triangles or quadrilaterals as the basic control-volume cells. This approach enables an easy and flexible treatment of general geometries. A development of the boundary conditions tailored for the current scheme is given. Fundamental validation tests are presented.

ISSN:1061-8562    

DOI:10.1080/10618569408904509

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

The purpose of this research is to numerically study a drag reduction method—passive control of shock/boundary layer interaction, which is applied to the boattail portion of a secant-ogive-cylinder-boattail projectile in turbulent transonic flows. The flow pattern and the components of aerodynamic drag computed from numerical data are analyzed. The effectiveness of this method is studied by varying the values of parameters such as porosity distribution, maximum porosity factor and size of porous region. The conditions for optimal drag reduction are investigated and reported. The present results show that the use of this passive control method can not only reduce the boattail drag but also the base drag, and results in an additional 8% total drag reduction compared to that without the passive control technique. This passive control method can be an effective approach for the design of high-performance projectiles in the transonic regime.

ISSN:1061-8562    

DOI:10.1080/10618569408904510

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

A new eigenvalue search procedure using the Chebyshev pseudospectral technique has been developed for the solution of the Orr-Sommerfeld equations. It can deal accurately and flexibly with various velocity profiles. The axisymmetric linear instability characteristics oflhe annular Hagen-Poiseuille Bow have been studied with this procedure. Neutral curves were obtained for 1 > R = R1/R0> 0.06 where Rlis the inner radius and R0is the outer radius. When R = 1 which is the plane Poiseuille flow, the corresponding critical Reynolds number is 5772.22 αr; = 1.0209. As R decreases to 0.06, the critical Reynolds number increases to 2.6 × 106and the wave speed decreases to 0.0573. The wavenumber first decreases gradually from a, = 1.0209 at R = 1.0 to αr, = 0.90 at R = 0.3. and then it increases rapidly to αr, = 1.73 at R = 0.06.

ISSN:1061-8562    

DOI:10.1080/10618569408904511

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Third-order upwind finite element solutions of natural convection in a square cavity with differentially heated side walls are presented. The third-order upwind finite element scheme is developed on the basis of the Petrov-Galerkin formulation and a mixed method. A turbulent approach such as the two-equation turbulence model is not considered in the basic differential equations, used on a wide range of the Rayleigh numbers governing the dynamics of natural convection in the square cavity. Numerical compulations of natural convection flows at the Rayleigh numbers ranging from 107to 1010are achieved, and the solutions are compare to other results. Finally, the effectiveness of the third-order upwind finite element scheme to be applied to computation of natural convection problems is discussed.

ISSN:1061-8562    

DOI:10.1080/10618569408904512

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Third order Compact Upwind Difference (CUD) Schemes are applied to compute steady-state solutions to the inviscid and viscous Burgers equation. A simple estimate shows and numerical experiments confirm, that the error in the maximum norm is inversely proportional to the square of (he number of grid points in the viscous layer. The discretization error of the viscous term appears to dominate the overall performance of the scheme in the convection-diffusion problem.

ISSN:1061-8562    

DOI:10.1080/10618569408904513

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

An investigation into the effect of inlet boundary conditions choice on the solution of the time-averaged Navier-Stokes equations together with a κ-ϵ turbulence model has been carried out for a two-dimensional internal ventilation flow. The study shows that for large ventilation areas, with associated velocity gradients, the assignment of turbulence parameters has a substantial effect on the solution.

ISSN:1061-8562    

DOI:10.1080/10618569408904514

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor