next up previous contents
Next: The -method Up: Time-discretization Previous: Time-discretization

Introduction

  After application of the space discretization of momentum and transport equations and the implementation of the boundary conditions as described in Section 7, the discretized equations in the time-domain read:

 

 

 

where and denote algebraic vectors containing the velocity and pressure unknowns in grid points, is the discrete scalar grid function and the total number of scalar unknowns is given by N. For two-equation turbulence models in the absence of other physical effects, N = 2. Furthermore, M is the diagonal matrix containing the value of in the centroids on the diagonal, D and G are the discretized divergence and gradient operators, S represents the space discretization of the convection and viscous stress tensors and is an operator involving the discretization of convection and diffusion of the scalar. In fact this term may also be non-linear, but in our program it is treated as if it is linear. The vector contains the volume forces and boundary values of the velocities and represents the source term with respect to , which is generally a function of and , and the boundary conditions. The extra source term result from the anti-diffusive parts as deferred corrections to the first order upwind approximation.
The time discretization is performed with a standard technique for the solution of ordinary differential equations. At this moment only one type of time-solver is present: the so-called method, i.e. a linear combination of the forward and backward Euler schemes.
Generalized method ??????? (Jos, Guus)


Tatiana Tijanova
Wed Mar 26 10:36:42 MET 1997