I developped the TriaxialCompressionEngine, precisely, to solve the problem that you explain here. This engine can still be made a lot more versatile. But, its purpose is to handle transitions between different loading paths, and to be easily configurable by editing the xml. If you edit the xml, and - for instance - change manually the value of sigma_iso, the packing will converge to a different confining pressure. If you play with other flags/values you can also fix some walls, or launch a compression test (or extension if the strain rate is set negative). If you want to simulate some loading paths and you think you can't with the current engine, it means you have to code some new features in TriaxialCompressionEngine source code (example : proportional strains test). In that case, I suggest that you tell me what you need to do and we decide how to implement it (which should be relatively easy based on what already exists). Note that, at the end, you (or me) can commit the change to svn.
Remark1 : TriaxialCompressionEngine can, by default, compress a packing and start a compression test when the assembly is dense and stable. You can run a first simulation in compression, then edit the intermediate file that is saved at the end of the compaction stage, change strainRate=x to strainRate=-x, and run the extension test. Remark2 : The save/edit/modify/load procedure is not exactly user-friendly... It would be usefull sometimes to change the loading path, or anything else, in real time, in one simulation. Vaçlav is developping something to do this from a command line. It probably works already. It means that we can completely avoid the load/save operations if we use this feature. Remark3 : Vaçlav is also working on the triaxial engines now for his PhD research, so he will probably participate in the future discussions about this. Janek is also starting to use those engines for simple compression (no lateral stress) on snow. I guess you just dived in at the right moment. :) Bruno Luc SIBILLE a écrit : > Hi, > > First, I am discovering Yade, so I think I will have a lot of questions... > > By looking the structure of the file called scene.xml (created, for > instance, by the generator TriaxialTest), I understand that in the same > file are defined the media (the bodies, here spheres) with the > geometrical and physical (or mechanical) properties and the simulation > that we want to perform (with is specific parameters). > > My problem is the following one: a given simulation is linked here to a > given media, so can we do different simulations from a unique given media > ? I give an example here after based on a triaxial test: > > I perform an isotropic compression, after I stop the simulation and save > the result. I don't go further because I want to do different things with > this consolidated granular assembly. > Then I want to do a compression and an extension both from the state of > the media after the isotropic compression. So I perform a first simulation > (the compression) by using the consolidated granular assembly (the result > saved), then I perform a second simulation (the extension) by using the > consolidated granular assembly (always the same result saved). > I imagine that I cannot do that with Yade since the kind of simulation is > saved in the scene.xml file and reused (here the simulation "isotropic > compression" is saved and reused whereas I want to perform a "compression" > or an "extension"). > > Am I right ? What would be the different steps (in few words) to perform > the simulations described above? > > Best, > > Luc > > _______________________________________________ > 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
