Thanks a lot, Ruochun, for the clear and fast answer, this is really helpful,
Giovanni Il giorno mercoledì 28 giugno 2023 alle 18:42:00 UTC+2 Ruochun Zhang ha scritto: > Hi Giovanni, > > 1. In general, we should still use the "true" material properties since > they have physics meanings. In my opinion, in DEM simulations, Young's > modulus, CoR, grain shape and such, should be obtained from the grains in > use and their material, not the other way around. However, if using true > properties is too numerically challenging or does not give some bulk > properties we wanted, or some physics properties are just not available, we > could relax/tune some of them. To start off, I suggest just hand-tune it: > Try reproducing the experiments that show friction angles, viscosity etc. > (one of the demos is the repose angle test, and maybe we should add a > flowrate demo too...) with DEM-Engine, then run them with your educated > estimation of material properties, then measure the bulk properties in > question (for example, see line 224 in the ConePenetration demo to see > how bulk density is measured). Change the grain properties then repeat the > process to understand how the bulk properties change w.r.t. them, until you > have satisfactory bulk properties. For the properties you are interested > in, Young's modulus and CoR and nu are probably minor, and *mu, rolling > resistance (Crr), grain mass and grain shape* are probably more > impactful. Just a suggestion: Even if you are using DEM particles much > larger than true powder grains, try to use clump particle shapes that sort > of resemble the powder grain shape (or at least somewhat deviant from pure > spheres), because the particles shape is probably very impactful but less > easy to alter between simulations. > > Of course, if you have many material properties to tune and many bulk > properties to match, it is possible to automate the process using data > science technologies. That would be a longer discussion. Usually educated > estimates and a few trial runs work well for relatively simple physics like > you are simulating. > > 2. If you just want to total force and total torque, then you can use > *tracker > *objects to get that information. You may look into the Centrifuge demo > line 167 to see how contact acceleration and angular acceleration are > extracted from an object (note it is acceleration, so to get force you need > to multiply that by masses, as in the demo). You *Track *an object, then > obtain information about the tracked object using the *tracker *handle. > Force information can only be obtained from the simulation, the > visualization tool can do nothing about it. > > However, if you would like all the forces pairs (maybe for example, when > you need to do structure deformation co-simulation), you have two choices. > One is that you write contact pairs to files. See an example in the > GRCPrep_Part1 demo line 167, *WriteContactFile *method. You will have the > locations and forces of all the contacts. You should extract the "SM" type > of contacts only, that is the sphere--mesh contacts. The second choice is > doing something like in the FlexibleMesh demo (it's a new demo and you > may need to check out the newest version), line 253, use a tracker to get > all force pairs concerning a tracked object. The benefit of that is you get > the locations and forces as C++ vectors directly, so you may use it > directly in your script for other purposes. It makes more sense to > visualize pairwise forces too (for example using Paraview Glyph), should > you need a force visualization like in one of DEME's frontpage GIFs. But > based on your description you probably just need a total torque. > > Let me know if there is anything else, > Ruochun > > On Wednesday, June 28, 2023 at 8:22:24 AM UTC-5 Giovanni Bianchi wrote: > >> Hello, >> >> I am a new user of Chrono and DEM-Engine, and I am trying to simulate a >> mixer for powders using DEM-Engine. However, I have some issues: >> >> 1) How can I calibrate the contact properties between particles to obtain >> specific bulk properties of the material? I have measured friction angle, >> viscosity, bulk density, and cohesion, but I don't know how to translate >> them into the contact properties needed for a DEM simulation, such as the >> Young module, coefficient of restitution, friction coefficient, and so on. >> >> 2) I am building my simulation starting from the DEMO_mixer, and I am >> using Paraview to visualize the results. Referring to this demo, I managed >> to see the particles and their velocity, but I would like to calculate the >> forces acting on the blades and the moment about the rotation axis. Do you >> have any suggestions on how to do it with Paraview or with any other >> software? >> >> Thanks, >> >> Giovanni >> >> >> -- You received this message because you are subscribed to the Google Groups "ProjectChrono" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/05bd1c4d-7409-4052-8632-060e7b565d09n%40googlegroups.com.
