A short presentation on the Explicit and Implicit time integration schemes which I use in my advanced LS-DYNA class is included here. Its a evolving document but the latest version is available for download.
Time Integration (PDF, 346 Kb)
Note: Fixed the corrupted PDF file. Sorry for the inconvenience.
Hi,
If you check the formula of sound speed in shells given on page 22 of this documents, this has term 1 v^2, i think it should have been 1-v^2.
Kindly cross check
Regards,
Susheel Khanna.
Hi Suri,
selective timestep?
I have already dyna 971 in use, how can I set up this feature?
Is it a method to accelerate a calculation with different types of mesh sizes?
That means I can speed up a model let say 90% dominated by 1 micro s and 10% 0.2 micro s timestep. No more obligation to run the entire model with 0.2 micro s. Am I right?
If yes, wonderfull.
Is it for both smp and mpp, or there are some restrictions?
Sorry for too much questions
Regards
Amir
Amir,
Selective mass-scaling can be invoked using the IMSCL parameter in *CONTROL_TIMESTEP. It is supported by both SMP and MPP.
Suri
Susheel,
Thanks for the correction. The new document reflects this (v1).
Suri
Hi Suri,
The pdf file is damaged, unable to open !
Regards,
Nihar
Hi Suri,
I didn´t see any difference with IMSCL and without.
Her is my input:
…
*CONTROL_TIMESTEP
$$ DTINIT TSSFAC ISDO TSLIMT DT2MS LCTM ERODE MSIST
0.000 0.900000 0 0.000 -2.000E-7
0.000 0 -11
*SET_PART_LIST_TITLE
selective_time_step
11
200075
…
Couls you please specify more the usage and impact of it.
Regards
Amir
Amir,
For relatively low mass-scaled timestep, both should perform the same. You could try using 1e-6 and study the differences between the two which should show the significance of the selective mass-scaling.
Suri Bala
Suri,
I tried to open this document (also save and then open) from the link provided.
However, I got the ADOBE error
“The file is damaged and could not be repaired”.
Is there a special way of reading this document?
Thanks
Akram