Hi, Václav: Currently I think using unbalanced force is the best way to know whether your system is close to equilibrium.
Using the "velocity norm" appears to be a direct quantity, however, the "velocities goes to zero, or kinetic energy goes to zero" does not equal to the equilibrium state. Consider a mass-spring vibration case, when the velocity is zero for the mass, the acceleration for the mass is the maximum. It might be similar to the simple problem when we study basic physics in high school. How to define an object is "static or equilibrium" (Newton's first law)? Then you would think delta(vel_norm)/t_p = 0, where t_p = a period of time, but it might be difficult to determine your t_p in DEM? Feng Chen Graduate Student Department of Civil and Environmental Engineering 223 Perkins Hall University of Tennessee, Knoxville, 37996 http://fchen3.googlepages.com/home -----Original Message----- From: [EMAIL PROTECTED] on behalf of Bruno Chareyre Sent: Tue 3/25/2008 5:44 AM To: [email protected] Subject: Re: [Yade-users] staticity quantification Hehehe. You don't like the "UnbalancedForce" parameter? ;-) It is exactly what you are looking for, and its already used the way you want to use it (switch between isotropic and axial compression) TriaxialCompressionEngine : if ( UnbalancedForce<=StabilityCriterion && abs((meanStress-sigma_iso)/sigma_iso)<0.005 ) { if(currentState==STATE_ISO_COMPACTION && autoCompressionActivation){ doStateTransition(STATE_ISO_UNLOADING); "UnbalancedForce" is a normalized sum of forces on all bodies. It can be a bit higher than 1 when the system is really far from equilibrium (e.g. at the begining of the isotropic compression when only peripheral spheres are pushed by boxes), but it is generally between 0 and 1. At static equilibrium, it is exactly equal to 0, which of course never happen. A reasonable value to consider something is stabilised could be between 1% and 0.1%. Bruno Václav Smilauer a écrit : > Hello, > > I would like to quantify how much has a simulation approached static > state (provided such state exists). Has someone done that? I thought of > summing velocity norms (but what should be the threshold here?) or > perhaps observing whether global stiffness is almost-constant. (Bruno, > you're expert on that, how would you do that?) > > The reson for that is to programatically stop simulation once desired > state is reached, without user interaction, so that I cun say: run that > until it is done. > > Regards, > > Vaclav > _______________________________________________ > Yade-users mailing list > [email protected] > https://lists.berlios.de/mailman/listinfo/yade-users > > -- _______________ Chareyre Bruno Maitre de conference Institut National Polytechnique de Grenoble Laboratoire 3S (Soils Solids Structures) - bureau E145 BP 53 - 38041, Grenoble cedex 9 - France Tél : 33 4 56 52 86 21 Fax : 33 4 76 82 70 43 ________________ _______________________________________________ Yade-users mailing list [email protected] https://lists.berlios.de/mailman/listinfo/yade-users
_______________________________________________ Yade-users mailing list [email protected] https://lists.berlios.de/mailman/listinfo/yade-users
