hi,there,

when I use projection method to solve the incompressible flow around a ideal plate without thickness(for some reasons,I use dimensional NSe!),I find a strange thing,as I change the incoming velocity or(and) the viscosity,and keep the same Res.number,the code donot give a same result,when running my code,I find,when enlarging the viscosity and velocity(the same Re.),iterations for the temperal speed needs many more times(same crit.for convergence),compared with that of small viscosity and small incoming velocity,and the later cannot capture velocity profile for boundary layer,why?

sorry,other questions,if you want to capture the velocity profile among the boundary layer correctly,at least how many grids should be filled among the boundary layer,according to the Re.number(e.g.Re=100)?and the distance for the nearest grid to the surface?

thank

(1). In the book of "boundary Layer Theory" by Schlichting, there is a chapter on the exact solution of laminar boundary layer. (2). It also has a section h.flow in the inlet length of a straight channel and a circular pipe, where there is a picture of the velocity distribution for laminar flow in the inlet section of a channel. You can use it to plan your mesh distribution from the wall, and from the inlet. (3). In the boundary layer over a plate, the mesh has to be stretched from the wall to the free stream, and from the inlet also. You should be able to get good solution, if you can put 30 mesh points from the wall to the edge of the boundary layer. You also need additional mesh points to cover the area from the edge to the free stream. So, I would say 50 points normal to the wall (carefully planned). (4). Follow the same principle for the mesh in the stream-wise direction from the inlet. There is a region near the inlet where the boundary layer is thin and grows rapidly, so you have to be careful there.

Thanks John,

what I do simulation is fluid flow around a ideal plate with limited length but without thickness,and the thickness of the boundary layer is developing,if I take 30 mesh grids in the boundary layer,where will those grids be filled,in the middle of the plate?

(1). The book by Schlichting also give you the boundary layer thickness as a function of the plate length and the Reynolds number. (eq. 2.2 from my book)(2). So, compute the boundary layer thickness at trailing edge, and put the 30 points between the wall and the edge of boundary layer. (3). You should also adjust the boundary layer thickness so that it is smaller near the leading edge. (4). In the stream-wise direction, put 30 point in the front 25% and 30 points in the rest of the plate. I am assuming that the flow is parallel to the plate.