Time Derivatives and Moving Grid



next up previous contents
Next: Boundary Conditions Up: 3.3.4 Transformation of Differential Previous: Gradient and Divergence

Time Derivatives and Moving Grid

With moving grids the time derivatives must also be transformed. Assume a time dependent function in the time dependent domain . Then, for the time derivative in domain we get (3.3-56) from the chain rule.

 

Here, the time derivative on the left side is at a fixed position in the computational space, i.e. at a given grid point. The time derivative at the right side is at a fixed position in the physical space, i.e. the that appears in the governing physical equations. The quantity is the speed of the grid points. The last term in (3.3-56) resembles a convective term and accounts for the motion of the grid. Hereafter we will use for the speed of the grid points.

With the time derivatives transformed, only time derivatives at fixed points in the computational domain will appear in the equations. Therefore, all computations can be done on the fixed grid in the domain without interpolation even though the grid points are in motion in the physical space.

The expression (3.3-56) for the transformation of the time derivatives was obtained by a chain rule expansion. To get a conservative form for the transformed time derivatives we apply the divergence theorem to an area element at time and to the same element at time , . Then making up a balance equation and proceeding to the limit , we obtain (3.3-57).  

The structure of (3.3-57) is easily grasped. Since is a measure for the area, the first term covers the change of the amount of quantity within a cell, whereas the second term covers the flux through the moving cell boundaries.

Again, it is straightforward to proof that (3.3-56) and (3.3-57) are mathematically identical.



next up previous contents
Next: Boundary Conditions Up: 3.3.4 Transformation of Differential Previous: Gradient and Divergence



Martin Stiftinger
Wed Oct 19 13:03:34 MET 1994