[MBDyn-users] Modal analysis

Pierangelo Masarati masarati at aero.polimi.it
Fri Oct 16 15:16:42 CEST 2009


Romuald NORET wrote:
> Hello,
> 
> I am very sorry, I didn't find myself this part of the manual before.
> 
> So I run eigen analysis, on MBDyn 1.3.7, some keywords as "lower frequency 
> limit" or "upper frequency limit" don't look to work the the following 
> synthax worked without crash :
> "eigenanalysis:
>         0.0,
>         output eigenvectors,
>         use lapack;"
> 
> But after the simulation, there isn't .m created and there is nothing 
> about eigenvectors (same when I ask the full matrice) into .out file :
> "# Derivatives solution step at time 0 performed in 0 iterations with 
> 9.61481e-12 error
> # Initial solution after dummy steps at time 2e-06 performed in 0 
> iterations with 9.61481e-11 error
> # Key for lines starting with "Step":
> # Step Time TStep NIter ResErr SolErr SolConv
> Step 0 3e-06 1e-06 0 9.61481e-11 0 0
> Step 1 0.010002 0.01 0 1.92296e-14 0 0
> Step 2 0.020002 0.01 0 9.61481e-15 0 0
> Step 3 0.030002 0.01 0 9.61481e-15 0 0
> Step 4 0.040002 0.01 0 9.61481e-15 0 0
> Step 5 0.050002 0.01 0 9.61481e-15 0 0
> Step 6 0.060002 0.01 0 9.61481e-15 0 0
> Step 7 0.070002 0.01 0 9.61481e-15 0 0
> Step 8 0.080002 0.01 0 9.61481e-15 0 0
> Step 9 0.090002 0.01 0 9.61481e-15 0 0
> Step 10 0.100002 0.01 0 9.61481e-15 0 0"
> 
> Maybe it comes from my MBDyn version out of date (1.3.7)?

No, the version should be fine (the keywords it does not recognize were 
introduced later, but they are not essential).  My guess is that your 
build of MBDyn did not detect lapack, and thus simply ignores the 
eigenanalysis statement (it should give a warning on the screen, 
though).   I'm not sure how you can easily check whether lapack was 
built-in.  Are you using UNIX or Windows?  In case of UNIX, you should 
be able, by running the command "nm" on the executable, to find whether 
it contains a call to lapack's routine dggev.

Cheers, p.


More information about the MBDyn-users mailing list