Stuck? 
support@mecway.com
support@mecway.com
Mecway
FEA
Victor
About
- Username
- Victor
- Joined
- Visits
- 4,723
- Last Active
- Roles
- Administrator
Comments
-
The internal solver ignores the Elastic option so it'll be the same as the default option.
-
Thanks for sharing that, @mmartin. Did you find it was necessary to have such small time steps, which are 1/10 of the time resolution of the seismic displacement function? It's taking too long to solve so can you spoil the ending - does the column c…
-
Thanks for confirming this, @traemand, and especially the detailed steps. I don't test on Linux with each release so it's good to hear it still works!
-
Oh, I see now. I was somehow looking at the wrong file before. Your explanation does sound like what happened. I'll see about adding a warning in a future version. Fixing it properly will be more difficult than Refine x2 because it has to keep track…
-
Could you clarify where the problems happen? I don't see any duplicated elements nor the beams merged to the solids or shells when I apply Refine custom to the whole model or to just the shells.
-
The convergence tolerance for compression-only support is no longer available. It converges when there are 0 changes in contact status at each node. Compression only support is really a poor-man's contact that I added because the CCX interface wasn…
-
Indeed a surprisingly simple solution. I didn't know bonded contact works all the way out to the circumference of the end of a beam element. Unfortunately, the internal solver only connects them at the node, so it won't contact the edges of the hole…
-
Solution -> New force and moment -> Axial force.
-
Loads & Constraints -> Transfer temperatures from solution. More details in Help -> Tutorials section 2.4 Thermal Stress.
-
That number is showing the instantaneous scale factor. I'm not sure why since it's kind of unreadable. The maximum should still be what you set.
-
That can happen when it fails generating the volume mesh but after generating the surface mesh which it always does first. With the surface mesh, can you then turn it into solids with Mesh tools -> Automesh 3D? That will also show you any gaps t…
-
@cwharpe, does dragging nodes with the mouse so they snap to another node do what you mean by FROM -> TO?
-
The surface integral tool should work correctly for diffuse radiation because it's really measuring the heat flux inside the elements at the location of the surface where it has a single direction at each point. Heat can't escape the surface in any …
-
Unfortunately, I can't find the test I used. It was probably a fully closed box with a solid element inside it. I think the problem was the view factor seemed to be wrong - maybe temperature not as I expected? It wasn't a failure to converge.
-
I probably won't make it configurable. The location is managed by .Net which also does things like transferring settings to updated installs of the software. Microsoft has somehow decided that's where applications should store their settings and I t…
-
No, sorry. Perhaps use a symlink?
-
I hope it's really solved. Perhaps the test I did was wrong but it left me with little confidence in the cavity radiation. You can set orthotropic in the Thermal tab of material properties.
-
Interpolating is to decouple the number of displayed colors from the number of defined colors. That way, you can change the number of colors, perhaps to reveal detail in a special case, or to highlight a zero-crossing without redefining them all. A…
-
If you set the "Number of colors" to 10 but have 20 colors specified, it keeps them distinct. I understand those other issues and they're on my list for future updates. Although this is tedious, you can get predictable view orientations using the…
-
Internal solver doesn't do cavity radiation or automatic time stepping. 1. I can't find the problem with temperature BC and radiation on a shell surface. Could you give me a more specific example? When I try it, it fixes the surface temperature as …
-
A couple of factors that affect it: The specified colors include the top and bottom values in the range but the display uses the color interpolated at the middle of each sub-interval. So in your screenshot, the displayed red is for the whole sub-in…
-
If you use Gmsh (see manual) instead of Netgen, most of the adjacent edges end up with matching meshes so you can then merge them with Mesh tools -> Merge nearby nodes. However, in this case, there are some small misalignments that would need to …
-
In both cases, the slave nodes appear to be following curve of the master face. At the bottom, that's concave so the (false) straight segments of the displayed edges cover the slave nodes, whereas on the side, it's convex so the curve that the slave…
-
Well that was quick. *TIE works if you specify 1mm for Position tolerance. I guess the automatic tolerance was too small so it wasn't reaching some of the nodes. I chose 1mm because that's similar to the node spacings.
-
@dculp Yes I'll investigate it but might not be able to come up with a solution quickly or that generalizes well. Unfortunately, the reason there are 3 options is that none of them are the best in every way. Elastic avoids the common and often rand…
-
This is does look horrifying and I don't know why it's not connecting all the nodes properly. You seem to have set everything up correctly. I tried reducing the elastic stiffnesses in case they were causing numerical error, and refining the hole sur…
-
That limitation is for the default option in bonded contact. I'm not sure how reliable *TIE is with it, but since it also uses MPCs, there's still a possibility it may fail sometimes. One reason it works here could be that the frictionless supports …
-
@Sebastianmaklary *TIE operates on the expanded elements as Sergio said which means you should align their faces, not their midsurfaces to eliminate the node moving, which can easily cause that nonpositive jacobian problem.
-
@Sebastianmaklary I'm currently replacing the shell elements with MITC type elements that are based on continuum mechanics instead of plate theory. They perform much better with curved shapes though they still have some idealizations like constant s…
-
You can replace both the *STEP definitions generated by Mecway with custom ones that include stress. Add a custom model definition to the CCX branch containing the following: *STEP *STATIC *CLOAD 6170,2,1 *END STEP *STEP,PERTURBATION *FREQUENCY 5 *N…
