QUESTION:
I am performing a thermal transient in Workbench. Convection is defined by using a command object (to define spatially varying convection with SF using tabular loading).
The calculation is correct - but Workbench does not report the mid node results at this boundary, leading to alternating solution: 0 - result - 0 - result as shown in the attached plot.
Postprocessing in classic works well.
Steady state works well in Workbench.
Why does this happen only in WB and only for a thermal transient?
ANSWER:
In ANSYS models having SOLID90s using the QUASI solver option, we suppress the mid nodes on any face with applied convections. That is because the QUASI solver uses only a diagonalized convection matrix which does not work well with 8-noded faces (only the consistent formulation does), hence we drop down to 4 nodes.
From the Element Manual for 9.0:
For transient solutions using the THOPT,QUASI option, the program removes the midside nodes from any face with a convection load. A temperature solution is not available for them. Do not use the midside nodes on these faces in constraint equations or with contact. If you use these faces for those situations, remove the midside nodes first. For the time being, the only workaround I can think of is to postprocess in ANSYS.
Note 1: supported WB convection does not have this issue because we apply convection via 152 elements which don't show the problem. If you don't issue the "sf,all,conv,%conve%,22" the results are ok.
Note 2: Feature request 49915 pending as of 10/20/2006
|
QUESTION: I am performing a thermal transient in Workbench. Convection is defined by using a command object (to define spatially varying convection with SF using tabular loading). The calculation is correct - but Workbench does not report the mid node results at this boundary, leading to alternating solution: 0 - result - 0 - result as shown in the attached plot. Postprocessing in classic works well. Steady state works well in Workbench. Why does this happen only in Mechanical and only for a thermal transient? ANSWER: In ANSYS models having SOLID90s using the QUASI solver option, we suppress the mid nodes on any face with applied convections. That is because the QUASI solver uses only a diagonalized convection matrix which does not work well with 8-noded faces (only the consistent formulation does), hence we drop down to 4 nodes. From the Element Manual for 9.0: For transient solutions using the THOPT,QUASI option, the program removes the midside nodes from any face with a convection load. A temperature solution is not available for them. Do not use the midside nodes on these faces in constraint equations or with contact. If you use these faces for those situations, remove the midside nodes first. For the time being, the only workaround I can think of is to postprocess in ANSYS. Note 1: supported WB convection does not have this issue because we apply convection via 152 elements which don't show the problem. If you don't issue the "sf,all,conv,%conve%,22" the results are ok. Note 2: Feature request 49915 pending as of 10/20/2006 See graphic below. <div style='float: left; padding: 20px 0px 30px 30px;'><img alt="[17E12FF4-E326-4C8F-A5B8-6AC6F9EBEE97.png goes here]" title="17E12FF4-E326-4C8F-A5B8-6AC6F9EBEE97.png" src="https://www1.ansys.com/onyxkb/solutions-cust/attach/17E12FF4-E326-4C8F-A5B8-6AC6F9EBEE97.png"><div style='text-align:center; padding:8px 0px 0px 0px;'><i>Attached Image: 17E12FF4-E326-4C8F-A5B8-6AC6F9EBEE97.png</i> |
||
|