Stuck? 
support@mecway.com
support@mecway.com
Mecway
FEA
Victor
About
- Username
- Victor
- Joined
- Visits
- 4,659
- Last Active
- Roles
- Administrator
Comments
-
I'm not sure about CCX specifically but for most FEA solvers, memory use isn't proportional to the number of equations. It goes up a bit faster than that since the equations are stored in a 2D matrix not a 1D list of values. But it's a sparse matrix…
-
That's an omission from the manual sorry. The table is only for the internal solver. So yes, you can use line3 tube elements with nonlinear 3D and CCX.
-
With *DLOAD, you can also write a user subroutine which could probably read the data from a file but I don't know anything about how to do that. If you need to recompile CCX, there are instructions in the file included in Mecway as well as various o…
-
Looking at *DLOAD, it seems like if you have a fluid model with node pressures, you can use PxNP to apply them to element faces. So maybe you could make a fake fluid model with all the pressures specified by your data, then use *DLOAD to map them on…
-
Just in case it helps, if you have smallish displacements, a simpler way to apply external moments to solids is with the moment load on a face selection. Mecway converts it to distributed forces for CCX.
-
Thanks. This seems to be a bug in Netgen which is generating a corrupt surface preview. I don't know why it's happening but for the next version it'll show an error message and continue to read what it can so you can see the area of the model that h…
-
Thanks for reporting this. When did the error appear? Immediately after opening the file, after meshing, at the start of meshing, or something else? Can you post more details that are shown with it, or better, the STEP file itself?
-
I guess that pretty much anything bought in the past 5-10 years will be OK. Mecway is not very GPU heavy - it uses it for simple rendering of triangles which isn't the bottleneck in the graphics.
-
The cutting plane slider on the right of the toolbar hides foreground elements. That's OK if your surface is flat but would need several adjustments to see all over a curved surface. If you use Mecway as the preprocessor then it carries components …
-
That's the nature of these problems - there are so many ways for it to go wrong, you can't be sure until you isolate it.
-
file for previous post
-
No, I only changed E for the Cplate material. I left K at 5e10. Here are the files I used if you want to run it again. I agree your results seem wrong. The plate's deflection strangely increases when its stiffness increases from E=1 GPa to E=10 GPa…
-
What does the convergence graph look like? Is generally decreasing or oscillating between identical values? If there's no graph yet, it might be stuck on the first iteration. It's possible that the model is big enough to cause the solver to use the…
-
The displacement is smaller with 1E10 as you'd expect. See attached picture. If K is changing the result, then it needs to be bigger. Make it as large as possible without causing it to fail or take too long to converge. You might have trouble with …
-
It works OK for me. I used this plate material: *MATERIAL,NAME=Cplate *ELASTIC 1.0000E+08,0.3 and it caused double the displacement at the middle of the plate, compared to the original plate material with E=1e9.
-
To be a bit more specific - in the *STEP section Sergio mentioned, use *BOUNDARY to specify a displacement in the X direction that moves the plate. Also remove the existing zero X *BOUNDARY constraint on it.
-
That's beautiful. Thanks for the post.
-
I think the problem is caused by an interaction between the extremely high stiffness elastic supports and the Rayleigh damping on the same elements. I solved it by reducing the elastic support stiffness on the surround to 10^12 Pa/m. I didn't notice…
-
disla: It looks like you have Mode superposition method turned on in Analysis settings. Without that, it doesn't have those errors. suktan: I haven't worked out what's wrong yet. If I reduce beta, the problem gets pushed back to later time steps bu…
-
Not with the internal solver, sorry. You can get it to skip time steps uniformly with Decimation number of time steps under Analysis settings. For example, set it to 10 to get only 10 uniformly spaced steps, which would include a few at the end.
-
It works OK for me. Can you attach the broken model to reproduce the problem?
-
A rule of thumb for slope of pressure-overclosure curve is (Young's modulus / thickness) * 10 That should converge OK, then you can increase it by a factor of 10 at a time if penetration is too high. Too low and you get large penetration. Too high…
-
The hole tool isn't very smart so you can't use it quite as freely as in CAD. There's no special process other than selecting a node and using the tool. But the node you select should be part of the surface of a solid or shell, not interior to a sol…
-
Thanks for the pictures but I somehow still don't understand, sorry. If it's merging too many nodes and damaging the elements, you should be able to use a smaller tolerance. For meshes where I have no feeling for the scale, I usually zoom in to wha…
-
Merge nearby nodes is supposed to work with 0 tolerance if the coordinates are identical. Maybe the last digit got changed when importing through .unv? To confirm they're identical in Mecway, select all the same-coordinates nodes and open Node coord…
-
No problem, I'll try to be a bit more specific. 1) No need to look in the .frd file by hand. CGX and Mecway can both display stress. I don't remember a lot about CGX but they should be alongside displacement in a menu somewhere. You can also open t…
-
Two reasons: 1) The thickness of the voice_coil part is different on the two models - 2 mm and 5 mm. That accounts for some of the difference but not all of it. To find this, I compared their volumes using Tools -> Volume and that showed the ful…
-
I can see stress (on the order of -1e4 at time 4.4e-5) and the rings do compress as they make contact as well as penetrating, which is normal. So at first glance that seems OK. Do you find there's no stress data in the .frd file or that it's too lo…
-
No, sorry. Though you can right click the graph and use Export to CSV file... then open it in Excel to graph it differently.
-
I see. Mode 1 is missing in the 1-segment model and both have the 279 Hz mode. This does look like a problem and it seems to be something to do with the frictionless support on the tri6 element at the center. I'll look into it further to see if it's…
