[pangxiongx]

Hello all,

I am doing mesh sensitivity studies of a rod bundle flow channel with star-ccm+. In the middle part of the channel there are mixing vanes - obstacles basically.

I divide the whole domain into three regions: region_1: "clean" channel upstream of the obstacles; region_2: channel containing the obstacles; and region_3: "clean" channel downstream of the obstacles.

For region_1 and 3, extruded (or directed) hexahedral mesh can be easily created; in region_2, since geometry is complicated because of obstacles, automatic trimmed mesh is used. The regions are connected with two interfaces.

My question is: while doing mesh sensitivity study in my case, can I do it separately on hexahedral and trimmed mesh, respectively? After the best mesh has been determined for both hexahedral and trimmed mesh, connect them with interfaces and then calculate the whole domain. Or at least after this process, in order to consider the wake flow downstream of obstacles, maybe I can only refine the mesh in region_3 in the axial direction (instead of 3 directions)?

Any hint or comment is appreciated. Thanks in advance!

Kind regards,
Pangx