[MBDyn-users] Output time period in *.log

Jens.vanSchelve at Emerson.com Jens.vanSchelve at Emerson.com
Wed Feb 16 18:32:16 CET 2011


 
You are right, but in this case it is difficult to link the time to the output data.
Therefore it might also be a solution to add some kind of output flag for each timestep in the *.out file.
E.g. just add one more collumn to the *.out file which indicates if this step is also passed to the output or not (simple 1-0 flag).

Jens

-----Ursprüngliche Nachricht-----
Von: Pierangelo Masarati [mailto:masarati at aero.polimi.it] 
Gesendet: Mittwoch, 16. Februar 2011 18:14
An: van Schelve, Jens [INDAUTO/SSB/SAL]
Cc: mbdyn-users at mbdyn.org
Betreff: Re: [MBDyn-users] Output time period in *.log

Jens.vanSchelve at Emerson.com wrote:
> Giuseppe: The *.out file stores information of EVERY timestep ignoring 
> if the data is passed to the output (*.mov, *.jnt...) files.
> So if the simulation itself starts a second x its ok, but not if only 
> the output starts at second x.
> 
> Pierangelo: So it would not be possible, too complicated or conflict 
> other softwares to add something like "output time: <starttime> 
> <endtime>" to the *.log file? -> Considering the fact that "output
> frequency: <frequency>" is already there? Or do you think of some 
> other special cases, because of the generec drive caller, that there 
> might be several start- and end times? Which would also make the 
> standart output files even harder to read.

I think what you ask is nearly impossible because of the generality of the drive caller.  In fact, the "output meter" statement has no notion of "start time" or "end time"; you could craft the output meter in a manner that output only occurs when a velocity passes a given threshold...  What should be easy to do is add an

output meter: <drive>

where <drive> is the echo of the drive caller.

> I was also considering the NetCDF output already, but our current 
> mbdyn post-processor is based on the standart output files and am not 
> quite sure if I find time to implement a MBDyn-NetCDF parser, however 
> I already looked into some NetCDF c++ examples and at some point in 
> the future I will do this.

In principle, the nice feature of NetCDF output is that you access the database using some high-level API; I usually use octave, but there seem to be quite a lot of tools that handle it.

Cheers, p.

> Having the start and end time of the
> output in the output files somewhere would just be much quicker to 
> implement for now :) It is still not that critical and was just meant 
> to reduce the amount of output.



More information about the MBDyn-users mailing list