[Afaraj]

Hello,
Please I need why when I switched from tetrahedron to polyhedron the number of cells reduced to five times compared to tetrahedron cell which I created them by ICEM. they were about 5.785 mi cells and reduced to 1.2 mi cells when I change meshing in fluent to polyhedral cells option.!!!?
Please help me on this question guys.

[ghost82]

A polyhedral cell is made of tetrahedral cells. When in fluent you switch to polyhedral tetrahedral cells are merged and this is because the total number of cells decreases.

[kaeran]

Considering specific applications converting from tetrahedral to polyhedral cells is always an advantageous in Fluent
This conversion reduces skewness and improves the mesh quality. So, you wouldn't be having any problem in convergence.
However, polyhedral meshes have disadvantages too- meshes will not be aligned in the flow direction.

Moreover reduction in mesh count is always realizable because more than one or two tetrahedral cells combines to form a polyhedral cells.
In my point of view, this is advantageous as you could reduce the simulation time. (in your case 5.7 million to 1.2 million)

[RodriguezFatz]

Originally Posted by kaeran

In my point of view, this is advantageous as you could reduce the simulation time. (in your case 5.7 million to 1.2 million)

But you also reduce the number of computational points and thus the accuracy, right? Or does one polyhedral cell have more than one point?
You can also reduce the number of points by just using a coarser tetrahedral mesh...

[kaeran]

RodriguezFatz: you are right -accuracy of the results will be affected, that's why I quoted disadvantages too.

For each cell either tetrahedral or quadrilateral or polyhedral, computation will be done considering their centroids.
For quadrilateral cells this centroid alignment is good whereas for tetrahedral it is slightly misaligned (but it depends on the domain either solid or fluid geometry).

In polygonal meshes, as you have more than 6 faces, centroid alignment will not be good which leads to non-realizable velocity or pressure profile calculations.
Anyhow, this type of meshing is useful in heat transfer calculations involving complex geometries.

However, if you experimental validation results yields good results with polygonal meshes, then you can go ahead.

[Afaraj]

A major advantage of polyhedral cells is that they have many neighbors
(typically of order 10), so gradients can be much better approximated
(using linear shape functions and the information from nearest
neighbors only) than is the case with tetrahedral cells. Even along wall
edges and at corners, a polyhedral cell is likely to have a couple of
neighbors, thus allowing for a reasonable prediction of both gradients
and local flow distribution. The fact that more neighbors means more
storage and computing operations per cell is more than compensated
by a higher accuracy, as will be demonstrated below.
Polyhedral cells are also less
sensitive to stretching than tetrahedra.
Smart grid generation and optimization techniques offer limitless
possibilities: cells can automatically be joined, split, or modified by
introducing additional points, edges and faces. Indeed, substantial
improvements in grid quality are expected in the future, benefiting both
solver efficiency and accuracy of solutions. Also, many situations with
tetrahedral meshes requiring special treatment in the solver cease to
be special when using polyhedral cell topology: cell-wise local mesh
refinement, sliding grid interfaces, periodic boundaries, etc., only
create special types of polyhedra, but for the solver they are all the
same.
Polyhedral cells are especially beneficial for handling recirculating
flows. Tests have shown that, for example, in the cubic lid-driven cavity
flow, many fewer polyhedra are needed to achieve a specified
accuracy than even Cartesian hexahedra (which one would expect to
be optimal for rectangular solution domains).
http://www.plmmarketplace.com/upload...polyhedral.pdf

[hellsblade.91]

In what kind of applications precisely does one use polyhedral mesh in Fluent and what are the advantages other than the fact being less of elements so less computational time