[ANSYS Meshing] meshing extremely slow

dynamic · February 13, 2017, 12:03

Hi,
I’m currently trying to mesh a three bladed VAWT with support struts, however this model has become extremely slow. Currently the mesher is taking approximately 60 hours to mesh a 5 million element model, previously when there was no struts the mesher was completing the meshing of a model containing 16 million elements in approximately 5 hours.

I’ve observed from windows task manager that under half my RAM is being utilized and only 39% CPU usage, is there any way I could increase these values to reduce my meshing time?

My set up is x64 windows 7 with 24 GB RAM. With an Intel i7-490 processor and I’m running the mesher on ANSYS 17.1

I’ve read elsewhere on this forum that changing the mechanical APDL values in workbench tools solves this problem, however when I change from the default 2 processors to 4, it has no effect on the meshing time.

Could someone please inform me of the correct method or share their expertise.

Looking forward to your replies.

Thanks and have a nice day 

vasava · February 14, 2017, 04:39

5 million cells with that computer should not take more than 20-30 mins.

Can you post details of sizing, size function and other setup?

dynamic · February 14, 2017, 04:59

Hi Vasava,
Thanks for your interest in my post.

Im using the patch conforming method (tetrahedrons) for both my static and rotating domains. I have 20 inflation layers over the blades struts and shaft. I have an edge sizing of 340 divisions over the blades at both ends which has a chord of 150 mm. I have edge sizing of 120 divisions over the shaft at both ends, as well as 240 divisions of the struts edges. and a division of 150 elements along the length of the blades of 1m
I also have an edge sizing on the edges of both domains totaling 4 (2 domains top and bottom) All these sizing have a "Hard" behavior.

Im also mapping the faces of theses, via the long faces. ie, the two lengthwise faces of the blades, the lengthwise faces of the struts and the length wise face of the shaft, aswell as the length wise faces of the 2 domains, I suspect it because of this mapped faced meshing that is slowing it right down.

is there anyway for the mesher to utilize more of the capabilities of my machine?

Thanks,

vasava · February 15, 2017, 05:11

Can you share your workbench files? Its more easier and faster than reading/imagining.

February 13, 2017, 12:03	meshing extremely slow	#1
dynamic New Member Ronald Thompson Join Date: Mar 2015 Location: Portland, Oregon, USA Posts: 25 Rep Power: 11	Hi, I’m currently trying to mesh a three bladed VAWT with support struts, however this model has become extremely slow. Currently the mesher is taking approximately 60 hours to mesh a 5 million element model, previously when there was no struts the mesher was completing the meshing of a model containing 16 million elements in approximately 5 hours. I’ve observed from windows task manager that under half my RAM is being utilized and only 39% CPU usage, is there any way I could increase these values to reduce my meshing time? My set up is x64 windows 7 with 24 GB RAM. With an Intel i7-490 processor and I’m running the mesher on ANSYS 17.1 I’ve read elsewhere on this forum that changing the mechanical APDL values in workbench tools solves this problem, however when I change from the default 2 processors to 4, it has no effect on the meshing time. Could someone please inform me of the correct method or share their expertise. Looking forward to your replies. Thanks and have a nice day 

February 14, 2017, 04:59		#3
dynamic New Member Ronald Thompson Join Date: Mar 2015 Location: Portland, Oregon, USA Posts: 25 Rep Power: 11	Hi Vasava, Thanks for your interest in my post. Im using the patch conforming method (tetrahedrons) for both my static and rotating domains. I have 20 inflation layers over the blades struts and shaft. I have an edge sizing of 340 divisions over the blades at both ends which has a chord of 150 mm. I have edge sizing of 120 divisions over the shaft at both ends, as well as 240 divisions of the struts edges. and a division of 150 elements along the length of the blades of 1m I also have an edge sizing on the edges of both domains totaling 4 (2 domains top and bottom) All these sizing have a "Hard" behavior. Im also mapping the faces of theses, via the long faces. ie, the two lengthwise faces of the blades, the lengthwise faces of the struts and the length wise face of the shaft, aswell as the length wise faces of the 2 domains, I suspect it because of this mapped faced meshing that is slowing it right down. is there anyway for the mesher to utilize more of the capabilities of my machine? Thanks, Last edited by dynamic; February 14, 2017 at 04:59. Reason: faces not edges

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[ANSYS Meshing] How to increase meshing speed?	Sylwia	ANSYS Meshing & Geometry	3	July 23, 2015 09:02
[ANSYS Meshing] Cooper Method Meshing	RTN3000	ANSYS Meshing & Geometry	1	March 1, 2015 23:59
Urgent help, help pl for meshing problem	Shashi	FLUENT	8	October 8, 2008 11:24
Meshing locks workbench window.	andy2o	CFX	0	February 1, 2008 05:01
Singularity of grid?Volume meshing vs face meshing	Ken	Main CFD Forum	0	September 4, 2003 11:09

February 14, 2017, 04:39		#2
vasava Senior Member Paritosh Vasava Join Date: Oct 2012 Location: Lappeenranta, Finland Posts: 732 Rep Power: 23	5 million cells with that computer should not take more than 20-30 mins. Can you post details of sizing, size function and other setup?

February 15, 2017, 05:11		#4
vasava Senior Member Paritosh Vasava Join Date: Oct 2012 Location: Lappeenranta, Finland Posts: 732 Rep Power: 23	Can you share your workbench files? Its more easier and faster than reading/imagining.