Don't go too far with code loading spheres, have a look at Shop::loadSpheresFromFile (just returns pairs of center-radius); and Shop::saveSpheresToFile. Vaclav > I > > > Václav S a écrit : > >>> 4. Watch the compaction stage >>> 5. See the spheres disapearing when compression is supposed toi start >>> >>> Also, I noticed that interaction types are SDECLinks in the triaxial >>> test, which is not supposed to happen. >>> >>> >>> >> Huh, that is my fault again. TriaxialCompressionEngine:86 should be >> if(0), not if(1); and the whole block can be effectively deleted. It was >> an experimental stuff to test what happens if I glue particles together >> after compaction. I just commited the fix. >> >> > Ok. Does it fix the SDECLink problem? > TriaxialTest.cpp was also using elasticContactLaw and > elasticCohesiveContactLaw at the same time, I don't know if you or me > commited that but I reverted that and will commit soon : TriaxialTest is > not supposed to be cohesive + 2 simultaneous contact laws makes no sense. > > >> Do you have some precise idea what the regression test will look like? >> Notably, what will be the reference value checked after the run. >> >> >> > I think the value of stress and strain after X iterations + the peak > stress value would make good tests. If they don't change you can be sure > that the physics and mathematics in Yade were not affected, because they > result from very complex processes, highly sensitive to initial > conditions (Lorentz sort of). > > The only thing is we need to swith from random positioning to a text > file, I still have to create this file and to get the code for loading > positions back. > > Bruno > > > > > >> V. >> >> _______________________________________________ >> yade-dev mailing list >> [email protected] >> https://lists.berlios.de/mailman/listinfo/yade-dev >> >> >> > > >
_______________________________________________ yade-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/yade-dev
