[rks171]

I can't seem to get a nice looking mesh around a cylinder I'm modeling. See picture below. The cylinder goes perpendicular to the screen. The flow is in the same direction. This is a cross-sectional shot of the cylinder in the flow field. I don't see why I can't get the prism layer to be nice consistent squares around the cylinder. Why does the trim mesh have so many non-square funky looking cells?




Here's the mesh portion of the summary report if that helps:

 | +-1 Mesh 1 	Verbose Output	false
  | | |  	OOC translation	false
  | | |  	Per-Region Meshing	false
  | | |  	Use Parallel Meshing	false
  | | |  	Interpolation Option	Nearest neighbor
  | | |  	Interfaces	[]
  | | |  	Regions	[Body 1]
  | | +-1 Models 	 	 
  | | | +-1 Prism Layer Mesher 	Stretching Function	Geometric Progression
  | | | |  	Stretching Mode	Stretch Factor
  | | | |  	Gap Fill Percentage	49.0
  | | | |  	Minimum Thickness Percentage	10.0
  | | | |  	Layer Reduction Percentage	10.0
  | | | |  	Boundary March Angle	50.0
  | | | |  	Concave Angle Limit	0.0
  | | | |  	Convex Angle Limit	360.0
  | | | |  	Near Core Layer Aspect Ratio	0.0
  | | | |  	Generate Standard Cells Only	false
  | | | |  	Improve Subsurface Quality	true
  | | | +-2 Surface Remesher 	Do curvature refinement	true
  | | | |  	Do proximity refinement	true
  | | | |  	Do compatibility refinement	false
  | | | |  	Retain geometric features	true
  | | | |  	Create aligned meshes	true
  | | | |  	Minimum face quality	0.05
  | | | |  	Enable automatic surface repair	true
  | | | `-3 Trimmer 	Coordinate System	Laboratory
  | | |    	Do mesh alignment	false
  | | |    	Template mesh type	Hexahedra
  | | |    	Template mesh growth type	Simple
  | | |    	Run Optimizer	true
  | | +-2 Reference Values 	 	 
  | | | +-1 Base Size 	Value	0.05 m
  | | | +-2 Automatic Surface Repair 	Connected surface count limit	None
  | | | | |  	Connected surface size limits	None
  | | | | +-1 Minimum Proximity 	Minimum Proximity	0.05
  | | | | `-2 Minimum Quality 	Minimum Quality	0.01
  | | | +-3 CAD Projection 	Project to CAD	true
  | | | +-4 Maximum Cell Size 	Size type	Relative to base
  | | | | `-1 Relative Size 	Percentage of Base	10000.0
  | | | |    	Absolute Size	5.0 m
  | | | +-5 Number of Prism Layers 	Number of Prism Layers	1
  | | | +-6 Prism Layer Stretching 	Prism Layer Stretching	1.5
  | | | +-7 Prism Layer Thickness 	Size type	Relative to base
  | | | | `-1 Relative Size 	Percentage of Base	1.0
  | | | |    	Absolute Size	5.0E-4 m
  | | | +-8 Surface Curvature 	Enable curvature deviation distance	false
  | | | | `-1 Basic Curvature 	# Pts/circle	36.0
  | | | +-9 Surface Growth Rate 	Surface Growth Rate	1.3
  | | | +-10 Surface Proximity 	# Points in gap	2.0
  | | | |  	Search Floor	0.0 m
  | | | +-11 Surface Size 	Relative/Absolute	Relative to base
  | | | | |  	Size Method	Min and Target
  | | | | +-1 Relative Minimum Size 	Percentage of Base	1.0
  | | | | |  	Absolute Size	5.0E-4 m
  | | | | `-2 Relative Target Size 	Percentage of Base	10.0
  | | | |    	Absolute Size	0.0050 m
  | | | `-12 Template Growth Rate 	Default Growth Rate	Medium
  | | |    	Boundary Growth Rate	Medium
  | | `-3 Volumetric Controls

[Paulh]

My guess is that you're cutting plane is located on a plane of cell faces. Try moving your section location just a bit. Your problem might be just a visualization issue.

[rks171]

I also thought that, but when I moved the plane, the mesh looks almost the same. I didn't mention that the cross-sectional area doesn't change at all. The cylinder and flow section dimensions are constant moving in the z-direction (in and out of the screen). So I don't see why mesh cell dimensions and cross-sectional area should change in the z-direction. At least, I don't want it to be changing in the z-direction.

[ryancoe]

Meshing a circular cylinder with cubes will inherently result in misshapen cells.

Alternatively, you could use a cylindrical mesh (gridgen and pointwise can do this, but Star-CCM+ cannot). Within Star-CCM+, you could try the polyhedral mesh...

[rks171]

I was able to improve this mesh simply by reducing the base size of my mesh cells from 0.05 m to 0.03 m. I played with a lot of other features, too, but that seems to be the one that made the funny looking cells go away around the cylinder.