CFD Online Discussion Forums

(1). It is not new or unique for the term "hybrid mesh". (2). Historically, if you take the unstructured tri/tet mesh to fill out the computational domain, it is nice and flexible for complex geometry problem. (3). But, pretty soon people realized that there is a physical side of the solution which needs to be considered as well. As a matter of fact, this physical side of the solution dictates the accuracy of the solution. (4). In other words, tri/tet mesh alone is not going to work! But why? (5). Because you are not looking for a smooth tri/tet mesh, you are looking for a physically correct solution, which includes the wall boundary layers, mexing layers, or discontinuities all the time. And the tri/tet mesh alone is bad in handling such features. (6). So, in order to make it work, it is necessary to cut a portion of the flow field next to the wall, or the narrrow region of the mixing layer, so that another type of mesh can be used for those regions. Normally, hex, or brick, or prism type are used in these high flow gradient region. (7). Since the mesh used is not just one pure type, there are more than one involved, it is called hybrid. (8). This type can happen in finite difference method, there, it is traditionally called multi-block method. You can have O-type mesh inside H-type mesh, or next to it. (9). In Fluent,from my point of view, "hybrid mesh" is something which is trying to recognize the existence of wall boundary layers, or difference in the geometry types. (10). In other words, tri/tet mesh alone is not going to give you the correct wall quantities, such as the skin friction and the heat transfer, unless you use very fine mesh. (it is very expensive in 2-D, and it is almost impossible in 3-D using tri/tet alone to resolve the boundary layers in complex flow problem.)