I need help in understanding the settings in NXSTRAT for solving solution 601,106.
First of all: I try to analyze a solid model of a cylindrical elastomer plate pressed into a rigid torus ring by use of a kind of spheric stamp connected with contact definitions between stamp and plate and plate and torus ring and vice versa.
As load I use an enforced displacement for the stamp in vertical direction.
The solution I did use is momentarily:
TLA-S Total load application with stabilization (solving in 10s by default) with
- Contact stabilization,
- Low-Speed Dynamics,
- Matrix stabilization.
The solution passes different buckling forms for the plate and thats what I expected.
Now my problem:
I 've got only 50 iteration steps after solving but I did put into NXSTRAT
TLANSTP = 400
50 is the default value and therefore I assume that my setting here is ignored or conflicting with something. I 've found a TSTEP definition from Case Control settings with 200 steps of 0.05s (also 10s total) in my dat-file.
From former solution attemps I also found parameters for ATS scheme and LDC scheme.
My questions therefore are:
1. Do these entries defining time characteristics conflict with each other?
2. Do the different option settings in NXSTRAT conflict with each other?
3. Do I have to clean up my strategy parameters in the solution every time completely?
4. What effect does such a missmash have?
The solution is somewhat time consuming and therefore I doe not want to find out with "TRY and ERROR".
Is there someone who can help me? - Any experiences are wellcome.
With best regards,
Her are my complete entries in DAT-File
ECHO = NONE
BCSET = 100
TSTEP = 100
LABEL = Subcase - Nonlinear Implicit
DLOAD = 301
SPC = 101
TSTEP 100 2000.050000 1
$* NX Modeling Object: TLA-S+ConStab
NXSTRAT 100 SOLVER 2 AUTO 4 NPOSIT 1 +
+ MSTAB 2 UPFORM 1 LSUPPER1.000000 MAXITE 25+
+ ATSSUBD 500 LDCDMAX300.0000 LDCCONT 1 TLANSTP 400+
+ TLAMXDF0.005000 NSUPP 10 LDCGRID 51086 LDCDOF 3+
I think you still need to turn on ATS. Changes in the settings of ATS or TLA sceme do not take effect until you choose on of it. You do that here:
As you can only choose on at a time, I think they won't get in conflic with each other.
made very good experience with ATS with its standard values. Only with ATSNEXT I had some problems, as often it ingores what I wanted!
For your question No.3 there is the possibility to have several "Strategy Parameter" Sets. You can name them, save them and then even use them in other solutions within the same sim-file.
I hope this helps a bit!
tank you for answering. - Here my remarks for you to to make it more clearly.
I. First, I tried ATS-Scheme in former attempts but it did not converge. Therefore I switched to TLA-S. Besides, I tried LDS in a long history of attempts, too.
I use different parameter settings already before in this model as you described it. And may be that is the problem.
I copied the "Strategy parameter"-Settings made for ATS-scheme into an new set for TLA-S-scheme and therefore the new parameter setting still contains values representing ATS-scheme in subset for ATS.
But I also specified new parameter values for TLA-S in the subset for TLA-S and was not worried about it.
But after checking my NXSTRAT entry IN MY EXPOERTED DAT-FILE I 'd found BOTH, parameters valid for TLA-S-scheme AND other parameters valid for ATS-scheme (and LDC-scheme) in the same NXSTRAT- entry as described earlier.
That's why I did ask for effects of parameter mishmatch in NXSTRAT.
Do I have to clean the ATS (and LDC) subset in my TLA-S strategy parameter setting or does NX NASTRAN only uses those parameters from NXSTRAT-Entry valid for TLA-S- solution and ignores the other ones?
II. And how can I correct my settings to get more than 50 iterations stored in TLA-S scheme? - Do I have to fit the numbers of time steps in TLANSTP of TLA-S scheme subset and in Time intervall settings under conditions of TLA-S behaviour in mapping complete soution time onto 10 seconds as described in the manual?
Which number of Time steps controls the number of iteration strored into the OP2-file?
Do you know this? - Or is it impossible to increase the number of time steps in TLA(-S)?
I hopefully wait for your answers and of course the recommendations of all the others, who deal with such solutions ;-)
With best regards