I know that this is more a groups about SW and calculus problem but this is the only place I have for an analysis of results (ccx_2.19, test case example segment.inp)
As I understand, the circumferencial sides should be cyclic symmetry’s connected by MPC, one as master, the another as slave. I have difficulties in understanding the result, hereby are the screenshoots for the first two modes with stress rr
and rz (I have set a polar transform). Should the master and the slave side’s values be equal? (or I have not taken the true datasets ? Should it be pstress) If anyone could me explain this this would by good.
What are the boundary conditions and how is this model loaded ?
sorry for the delay. I have forgotten to write it is a pure frequency calculation : there no loads not fixed boundary. The input files are included in any CalculiX examples distribution. The image I had attached was a stress contour for the first mode computed. In the meanwhile I think I have understood better it. The frequency calculation outputs a stress contour which represents the maximum amplitude of a wave , another dataset would give a phase shift between points, am I correct ?
I have another problem : I have calculated the whole disk (always frequency step)
and I have discontinuous fields for some modes. The image below is for mode 6 (The max value of the contour is 2.54 E+07 and minilmal value is -2.54 E+07
¿How do the deformed shape look like?. ¿Are those nodes in the “discontinuity” connected?
What does that thick vertical line mean ? Isn’t that a disconnected mesh ? How was the full model created ?
You are right, there was spurious nodes. I corrected this by adding
merg n disk
In my pre.fbd file
Thanks a lot
Regarding this model, I have loaded the segment.inp
I’m computing the first 70 modes taking advantage of the cyclic symmetry constrain and comparing with the modes obtained with the full model.
The Cyclic constrain misses many modes that the full model is capturing.
¿Is this a known limitation or am I setting the model wrong?
With a sliced part you will only be able to find modes which are symmetrical within the size of the sliced part
Mmh. I don’t think so.
Cyclic symmetry is supposed to fix that. You are describing the ordinary symmetry BC.The attached modes would not be possible with just a 30º sector of the model. 129845.9Hz is the first mode.
I have updated the file comparing the sector with the full model.
Number of missing modes depends on the number of modes requested.
A minimum of 4/6 seems recommended for this model when *CYCLIC SYMMETRY MODEL and *FREQUENCY is combined.
Particular attention to the two first columns where 2nd and 4th modes are missing.
So far you have calculated on a full disc and one twelfth of a disc. If you try to calculate one seventeenth of a disc, then you will loss some more frequencies
Just tested and it provides the full set of modes (I have request 8 modes in this case). The missing modes depends on the number of modes requested not the size of the sector. *CYCLIC SYMMETRY MODEL doesn’t have any problem to extract modes which are not symmetrical within the size of the sliced part. That is a problem associated with the ordinary symmtry BC.