| John C. Chien
|
December 23, 2000 12:40 |
Re: turbulent mixing layers
(1). Yes. It is useless to use short paragraphs, if one can not get the message through. It is better to state the problem three times in three different angles. (2). Before the mixing, there must be a plate separating the two stream, therefore, you must take into account the existence of the dividing wall. (3). When the two streams meet at the trailing edge of the dividing plate, the flow field can be rather complex due to the geometry of the plate at the trailing edge. This is important, because it will impact on the solution. (4). Now ahead of the trailing edge, you have two separate streams or flow over a plate (wall), so, each stream will have a wall condition. (5). An at the place where the initial condition is known, you can set the initial condition there for each stream. And this will be the initial condition for your problem. You can specify the initial (inlet) condition of each stream as a uniform condiiton, or non-uniform condition (specify the profile). (6). As for the downstream condition, a simple parallel flow condition is adequate. (7) In other words, you will have a large box with a dividing plate in the middle, extending from the inlet location into the computational domain for a certain distance. That dividing plate will have an upper and lower wall, including the trailing edge geometry. The upper and the low outer boundary can be set to the corresponding inlet conditions. ( you can run a few cases with different domain sizes to check the effect of the outer boundary conditions) (8). This is a fairly standard problem, but you must start the computation before the two streams meet.
|
