Re: [Yade-users] [Question #452513]: overlapping particle/particle and particle/wall
Question #452513 on Yade changed: https://answers.launchpad.net/yade/+question/452513 Bruno Chareyre proposed the following answer: >"is there a way to minimize the overlap?" You can decrease confining pressure or increase material stiffness. The result will be the same both ways: less overlap, more cpu time. I am actually wondering if it is not helping to have overlaps for the fluent part of the problem. When the spheres are exactly tangent to the walls it is horribly demanding in terms of mesh refinement if you want to capture what happens near the tangent point (in fact a lubrication problem). Just curious. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #452513]: overlapping particle/particle and particle/wall
Question #452513 on Yade changed: https://answers.launchpad.net/yade/+question/452513 Bruno Chareyre proposed the following answer: "decrease confining pressure" should be "decrease gravitational acceleration" in this particular case. I just realized. B -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #452501]: Status of DFNflow
Question #452501 on Yade changed: https://answers.launchpad.net/yade/+question/452501 Status: Open => Answered Bruno Chareyre proposed the following answer: Hello, it is now implemented and functional. This is maybe why it looks like "stopped". I would suggest improvements/changes (if needed) instead of re-starting from scratch. Changing the functional dependency between aperture and conductivity (for instance) is not a big deal in itself, the main difficulties were much more algorithmic. The DFN code was developed as part of the PhD of Timos Papachristos, who is actually defending today (in 20min). Interesting synchronicity! Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #449025]: Capillary model
Question #449025 on Yade changed: https://answers.launchpad.net/yade/+question/449025 Bruno Chareyre proposed the following answer: Hi, What Jerôme suggests can be done in a script actually, i.e. adapting the deformation rate to the current stress (with user defined stress calculated in the script and including the capillary forces). B -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #446267]: Energy tracking of a two-sphere contact
Question #446267 on Yade changed: https://answers.launchpad.net/yade/+question/446267 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi Tijan, I would think 1) is as expected. When adhesion disappears there is still frictional resistance, the elastic force in excess wrt the maximum frictional force is counted as part of frictional dissipation (in the same step when the bond is broken). It should be substracted from bond breaking energy probably, to not count it twice. 2) is more a problem. The mismatch comes from the fact that the elastic potential is not exactly the sum of normEnergy[i] = 0.5*(i.phys.normalForce.squaredNorm()/i.phys.kn) shearEnergy[i] = 0.5*(i.phys.shearForce.squaredNorm()/i.phys.ks) There is coupling term combining the normal and shear components, which is small as long as the normal displacement is small compared to diameter, but which is there. This is what you see (since un is non zero and changing during your scenario). So I would say: the energy is conserved (*), but the internal energy you are plotting is only an approximation of the exact one. (*) And you can check that by integrating the external work along complex paths and find it path independant (without bond breakage). I happen to have a paper in preparation about this, I can send it to you later. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #432848]: yade compiled installation in a redhat cluster
Question #432848 on Yade changed: https://answers.launchpad.net/yade/+question/432848 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, You are probably not running the debug build if you don't see more output. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #439162]: KeyError: "Unknown energy name 'kineticParticle'
Question #439162 on Yade changed: https://answers.launchpad.net/yade/+question/439162 Bruno Chareyre posted a new comment: If I understand the above correctly you are copy/pasting lines selectively from one version to another. Then you end-up with a version that only you have. Good for you if it works this time, but in the long run it will be impossible to manage. If you are writing your own pieces of code I would suggest to keep using the latest version (i.e. including frequent updates) of the sources to avoid trouble. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #429604]: Cylinder and periodic boundary conditions
Question #429604 on Yade changed: https://answers.launchpad.net/yade/+question/429604 Status: Open => Answered Bruno Chareyre proposed the following answer: > How do you get the normal component of the relative velocity? Something like this (it would need another term from the velocity gradient, you can get inspiration from scgeom code): (state2->vel - state1->vel).dot(normal) >from a physical point of view I believe that the question of the contact point is not that simple and is fundamental. Of course it is not really half-way between non-symmetric objects (different sizes, different properties, ...) but the real question is where should it be if it is not half-way? Fundamentally, in fact, the notion of contact point/force is in itself irrelevant, that is a reason why I don't like very much the grammar " c1x = (geom.contactPoint - de1.pos)". I prefer to write equations without contact points, even if of course it can be mathematically equivalent. The difference between the two sets of equation is that the first one uses positions, the second doesn't. The problem in the first one is this (I didn't realize immediatly): if one body is larger than one period it is not clear how to define I->cellDist since it is not unique. If positions are not used the problem disappear. > but this would mean keeping all the compatibility with the existing ViscElPhys_Basic law (prescription possible of (young, poisson) or (kn, ks) with (en, et) or (cn, cs)) "Merging" A and B does not mean to break B to make it fit in A by brute force. ;) Of course the point is to keep the good stuff, else no point merging. I guess the straight way is to make a viscous version inheriting from CundallStrack. It may need to reorganize a little bit the code of CundallStrack to let sections of the go() function be used as standalone functions (mostly a way to avoid code duplication in the inheriting classes). Not sure yet if the law functor is enough or if some Ig/Ip functors would need something as well. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #434659]: the displacement diagram using periodic boundary
Question #434659 on Yade changed: https://answers.launchpad.net/yade/+question/434659 Status: Open => Answered Bruno Chareyre proposed the following answer: >how does the periodic boundary apply load in yade By changing the geometry of the periodic space. Arrange a long (infinite) chain of spheres along a line, with initially 10 spheres/meter and the same patern of spheres repeated every meter, then at the next step you decide that 1m becomes 0.9m, there is now 10spheres/0.9meter - they are squeezed and repulsive forces appear: you exerted a load. This would be 1D periodicity. And then, how can we plot this on a screen? We don't want to display and infinite number of spheres. So let's pick one randomly, call it "origin", and display the movement (relative to origin) of the next 10 spheres on its right hand side. It will look like they all go to the right in extension, and to the left in compression. That's really only a graphical display issue, not a mechanical question. By the way, if you compress something between two plates do you agree that it is strictly the same (in statics) to 1/ move the bottom plate by h and keep the top fixed, or 2/ move top by -h with fixed bottom, or 3/ move both by +/- h/2? Did you see the documentation btw? (theoretical background chapter) >and how does it choose the origin of coordinate? A puzzling question actually! The origin is placed at the point where the displacement is null. Where is it that the displacement is null? You may ask. Well... it is at the origin. Or are you asking the distance between the sun and the origin of the periodic space? ;) Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #429604]: Cylinder and periodic boundary conditions
Question #429604 on Yade changed: https://answers.launchpad.net/yade/+question/429604 Status: Open => Answered Bruno Chareyre proposed the following answer: > I also want the normal one Very easy to find. Will you really waste cpu time in calculating the incident shear twice when you actually only need the normal one (you already have the shear one with my suggestion). >contact point is always considered to be situated at the middle of the penetration depth. While this is ok for particles of the >same size, this to my opinion does not apply when considering the interaction of a particle with a wall. My question was (and >still is): why is it necessary to use this trick/approximation to avoid the bug ? I'm not following you fully. 1/ I don't see a trick/approximation, to me it is just ordinary code. So when you ask "why" I tend to just think "why not?". 2/ why do you think it is irrelevant to include the 0.5*un term in the branch between e.g. sphere and box or spheres of different sizes? How would you like it? Did you realize that the same 0.5 is also used between facets and spheres whatever the law functor? > I believe it is not a good idea to leave this contact law with a mistake You are right there is nothing wrong in fixing existing code and you are very welcome to do it. However, guess why ViscElPhys_Basic has a bug that CundallStrack does not have? Because they are two different functors instead of one, i.e. more lines of code with less eyes/fingers for each of them. Hence if someone wants to do something very useful for the years to come he could merge some of the existing functors. That was my point. Cheers Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #434659]: the displacement diagram using periodic boundary
Question #434659 on Yade changed: https://answers.launchpad.net/yade/+question/434659 Status: Open => Answered Bruno Chareyre proposed the following answer: >coordinate as (0,0,0) and use the maximum coordinate as (1,1,1), even if the values of minimum coordiante and maximum >coordinate in the absolute coordinate system are changing. Unfortunately no, that's not what I meant, but I understand it was unclear. The minimum is always (0,0,0) but the maximum can be anything and it is changing with time (even in the negative quadran after some time if the velocity gradient is rotational). Some implementations are indeed setting all coordinates between 0 and 1 as you suggest, but yade is not. When writing (1,1,1) I was mentally thinking (1*v1+1*v2+1*v3) where the vi's are the vectors defining the periodic box (and they are changing with time), it is not represented in this way in the code. >if I want to display the characteristic of displacement tendency under a certain load(such as pure shear), I should look >symmetric in the window from (-0.5,-0.5,-0.5) to (0.5,0.5,0.5). I'm not sure what "characteristic of displacement tendency" means, but if you ask how to see it symmetric, yes, you need to look through a window between -0.5*(v+v2+v3) and +0.5*(v+v2+v3). Which is technically impossible with yade without considerable changes in the source code, by the way. So, by default, simple shear will look like [1]. If you want to see it (skew)symmetric you superimpose a constant rigidbody velocity, in the post processing as Jan suggested (by far the simplest) or even in the simulation itself by assigning the same initial velocity to all particles. Bruno http://www.flowillustrator.com/wp- content/uploads/2015/02/ViscosityShearFlow.png -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #434659]: the displacement diagram using periodic boundary
Question #434659 on Yade changed: https://answers.launchpad.net/yade/+question/434659 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, I don't think you need evidence about that for writting a paper: the problem you are pointing out simply does not exist, so you better don't mention it. ;) A periodic problem is infinite in size and you are seeing of it is through a small window. Depending on where you put the window you see a different average (rigid body) displacement. In yade we look at a 3D window with corners (0,0,0) and (1,1,1), the problem would look symmetric in the window from (-0.5,-0.5,-0.5) to (0.5,0.5,0.5) but it would still be the same problem. That is all. In this context adding a constant disp is equivalent to changing the origin of the reference frame, it does not change the problem itself. It is of course acceptable to choose the origin as you like and it if arrows look nicer, it does not need to be commented in a paper. I hope it helps Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #429604]: Cylinder and periodic boundary conditions
Question #429604 on Yade changed: https://answers.launchpad.net/yade/+question/429604 Status: Open => Answered Bruno Chareyre proposed the following answer: My Christmas gift below. :) The cylinders are ok in periodic boundary conditions overall. It has been applied to sheared fiber suspensions for instance, where basically every fiber can have different nodes in different periods (if we admit that "being in different periods" makes any sense...). François also generated periodic grids IIRC. However having bodies larger than period is the door open to remaining bugs in interaction detection, that's most likely the source of this new problem. If you create a cylinder with length equal to period and discretized in three segments then none of the segments are too large and everything is ok. And you can still use the "largeBody" trick to insert the floor. Bruno __ from yade import pack from yade.gridpfacet import * dp= 6e-3 periodic = 1 cylinderBottom = 0 cylindersBottom = 1 ground = 0.2*dp H = 60*dp L = 20*dp O.engines = [ ForceResetter(), InsertionSortCollider([Bo1_Sphere_Aabb(),Bo1_Box_Aabb(),Bo1_GridConnection_Aabb()],allowBiggerThanPeriod = True), InteractionLoop( [Ig2_Sphere_Sphere_ScGeom(),Ig2_Box_Sphere_ScGeom(),Ig2_Sphere_GridConnection_ScGridCoGeom()], [Ip2_ViscElMat_ViscElMat_ViscElPhys()], [Law2_ScGeom_ViscElPhys_Basic()] # [Ip2_FrictMat_FrictMat_FrictPhys()], # [Law2_ScGeom_FrictPhys_CundallStrack()] ), PyRunner(command = 'check()', virtPeriod = 0.1), NewtonIntegrator(damping=0., gravity = (0,0,-9.81)) ] #Material creation O.materials.append(ViscElMat(en=0.5, et=1., young=5e6, poisson=0.5, density=2500., frictionAngle=0.4, label='Mat')) #O.materials.append(FrictMat(young = 5e6, poisson = 0.5, density=2500,frictionAngle=0.4, label='Mat')) #Periodic Cell or containing box if periodic: O.periodic = True O.cell.setBox(L,L,H) O.bodies.append(box(center= (L/2.,L/2.,ground),extents=(100,100,0),fixed=True,color = (0.,1.,0.),wire = True,material = 'Mat')) else: O.bodies.append(box(center= (L/2.,L/2.,ground),extents=(100,100,0),wire = True,fixed=True,color = (0.,1.,0.),material = 'Mat'))#Made invisible to see below O.bodies.append(box(center= (0,L/2.,H/2.),extents=(0,L/2.,H/2.),wire = True,fixed=True,color = (0.,1.,0.),material = 'Mat'))#Made invisible to see O.bodies.append(box(center= (L,L/2.,H/2.),extents=(0,L/2.,H/2.),fixed=True,color = (0.,1.,0.),material = 'Mat')) O.bodies.append(box(center= (L/2.,0,H/2.),extents=(L/2.,0,H/2.),wire = True,fixed=True,color = (0.,1.,0.),material = 'Mat'))#Made invisible to see O.bodies.append(box(center= (L/2.,L,H/2.),extents=(L/2.,0,H/2.),fixed=True,color = (0.,1.,0.),material = 'Mat')) #Bottom fixed cylinder or fixed particles if cylinderBottom: n = len(O.bodies) cylinder(begin = (L/2.,100*L+0.001*dp,ground+dp/2.),end =(L/2.,-100*L+0.001*dp,ground+dp/2.),radius = dp/2.,fixed = True,color = (0,0,1),intMaterial = 'Mat',extMaterial = 'Mat')#Made invisible to see inside else: if cylindersBottom: ep=0.001*dp #a small thing nodes=[[L/2.,ep,ground+dp/2.],[L/2.,0.33*L+ep,ground+dp/2.],[L/2.,0.33*2*L+ep,ground+dp/2.],[L/2.,L+ep,ground+dp/2.]] cylinderConnection([[L/2.,ep,ground+dp/2.],[L/2.,0.33*L+ep,ground+dp/2.],[L/2.,0.33*2*L+ep,ground+dp/2.],[L/2.,L+ep,ground+dp/2.]],dp/2.,fixed = True) else: for n in range(0,int(L/dp)+1): O.bodies.append(sphere(center = (L/2.,n*dp,ground+dp/2.),radius = dp/2.,fixed = True,color = (0,0,1),wire = True,material = 'Mat'))#Made invisible to see inside #Particle cloud for gravity deposition partCloud = pack.SpherePack() partCloud.makeCloud(minCorner=(4*L/10.,4*L/10.,ground+3.*dp),maxCorner=(6*L/10.,6*L/10.,0.9*H),rRelFuzz=0., rMean=dp/2., num = 200) partCloud.toSimulation(material='Mat') O.dt = 1e-5 O.saveTmp() #Function to check if the center of a particle is contained inside the cylinder or pseudo cylinder def check(): for b in O.bodies: if b.dynamic and b.state.pos[2]<ground+dp: posCylXZ = Vector3(L/2.,0.,ground+dp/2.) pRelX = b.state.pos[0]%L - posCylXZ[0] pRelZ = b.state.pos[2] - posCylXZ[2] if sqrt(pow(pRelX,2) + pow(pRelZ,2))https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #406523]: Cylinders and law2_ScGeom_ViscElPhys_Basic
Question #406523 on Yade changed: https://answers.launchpad.net/yade/+question/406523 Bruno Chareyre posted a new comment: Sounds good. :) -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #406523]: Cylinders and law2_ScGeom_ViscElPhys_Basic
Question #406523 on Yade changed:
https://answers.launchpad.net/yade/+question/406523
Status: Open => Answered
Bruno Chareyre proposed the following answer:
I did not mean that a cylinder was "made of" spheres, a cylinder IS a
sphere.
Lets define a sphere as something which has position and radius.
A cylinder (gridConnexion) has position + radius + length, hence it is a sphere
according to the above definition.
Programatically at least it is really the case:
class GridConnection: public Sphere{
...}
It is also true for the gridNodes, they are spheres as well for the same
reason.
So if no functor is found to handle interactions sphere+connexion, node-
node, etc. then the dispatcher is looking for a possible - more generic
- sphere+sphere functor. That's what happens in the script.
Conclusion: there is (should be) no problem for sphere-cylinder interactions
because they are actually sphere-sphere interactions.
I really don't know why Newton-Raphson is not converging for the viscous
node-node interaction (I just checked, the mass are not null), but I can tell
it is not a problem of functor availability for specific shapes. Actually you
couldn't care less: you want rigid cylinders so you don't care about the
node-node behaviour. Just fix all degrees of freedom and it should be fine.
Bruno
--
You received this question notification because your team yade-users is
an answer contact for Yade.
___
Mailing list: https://launchpad.net/~yade-users
Post to : yade-users@lists.launchpad.net
Unsubscribe : https://launchpad.net/~yade-users
More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #406523]: Cylinders and law2_ScGeom_ViscElPhys_Basic
Question #406523 on Yade changed: https://answers.launchpad.net/yade/+question/406523 Bruno Chareyre posted a new comment: Hi Raphael, Quick answer to your last point: it is not a problem to "evaluate the kn, ks, cn and cs from the young, poisson, en and es" between spheres and cylinders since by design (inheritance) the cylinders _are_ spheres. So the problem must be simply that the cylinder has a null mass or something, not that it is not a sphere. What happens between walls/boxes/facets and spheres by the way? They usually have null mass as well. Cheers Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #406523]: Cylinders and law2_ScGeom_ViscElPhys_Basic
Question #406523 on Yade changed: https://answers.launchpad.net/yade/+question/406523 Status: Open => Answered Bruno Chareyre proposed the following answer: A cause of the trouble is that the function cylinder(begin = (0,0,0),...) is by itself creating an interaction between the nodes 0 and 1. It uses "createInteraction(id1,id2)", which in principle checks the available functors in the simulation to determine the type. The problem is you create the cylinder before defining the functors, hence some default functors (FrictPhys) present at startup are used. Before any iteration you get: Yade [1]: i=O.interactions.nth(0) Yade [2]: i.id1,i.id2 -> [2]: (0, 1) Yade [3]: i.geom,i.phys -> [3]: (, ) If you start your script with O.engines=[] (removing the default functors) the result is different. Now the problem is to create the initial interaction 0+1: File "raphael.py", line 6, in cylinder(begin = (0,0,0),end =(0,10,0),radius = 1,fixed = True,color = (0,0,1),intMaterial = 'Mat',extMaterial = 'Mat') File "/usr/lib/x86_64-linux-gnu/yadedaily/py/yade/gridpfacet.py", line 167, in cylinder cylIds.append(O.bodies.append( gridConnection(id1,id2,radius=radius,wire=wire,color=color,highlight=highlight,material=extMaterial,mask=mask,cellDist=None) )) File "/usr/lib/x86_64-linux-gnu/yadedaily/py/yade/gridpfacet.py", line 110, in gridConnection i=createInteraction(id1,id2) RuntimeError: No IGeomDispatcher in engines or inside InteractionLoop. If you define the cylinder after the engines it seems to work as expected (not quite sure what happens with the damping though, you'll need to check), and there is no need to include Ig2_GridNode_GridNode_GridNodeGeom6D() since you don't care about none-node interactions for rigid cylinders. Try the script below and let me know if it is still strange. Cheers Bruno from yade.gridpfacet import * O.materials.append(ViscElMat(en=0.5, et=0., young=1e7, poisson=0.5, density=2500, frictionAngle=0.4, label='Mat')) O.engines = [ ForceResetter(), InsertionSortCollider([Bo1_Sphere_Aabb(),Bo1_GridConnection_Aabb(),]), InteractionLoop( [Ig2_Sphere_Sphere_ScGeom(),Ig2_Sphere_GridConnection_ScGridCoGeom()], [Ip2_ViscElMat_ViscElMat_ViscElPhys()], [Law2_ScGeom_ViscElPhys_Basic()], ), NewtonIntegrator(damping=0., gravity = (0,0,-9.81)) ] cylinder(begin = (0,0,0),end =(0,10,0),radius = 1,fixed = True,color = (0,0,1),intMaterial = 'Mat',extMaterial = 'Mat') O.bodies.append(sphere(center = (0,5,5),radius = 1,material = 'Mat')) O.dt = 1e-6 -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #406523]: Cylinders and law2_ScGeom_ViscElPhys_Basic
Question #406523 on Yade changed: https://answers.launchpad.net/yade/+question/406523 Status: Open => Needs information Bruno Chareyre requested more information: Very strange... Yade [2]: for b in O.bodies: ...: print b.material ...: Did you find which object is of FrictPhys type? -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #406578]: TesselationWrapper for 2d case
Question #406578 on Yade changed: https://answers.launchpad.net/yade/+question/406578 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi Jiaying, I never tried that specifically but I don't anticipate special difficulties. Did you find a problem? Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #406523]: Cylinders and law2_ScGeom_ViscElPhys_Basic
Question #406523 on Yade changed: https://answers.launchpad.net/yade/+question/406523 Status: Open => Answered Bruno Chareyre proposed the following answer: >Am I right if I say that cylinders are not independant objects and are made of gridConnection and gridNode ? If yes, is there another element entering inside the construction ? Yes, correct. And no other element. >For the bounding box, I have two possibilities: Bo1_GridConnection_Aabb() or Bo1_Cylinder_Aabb() No, Bo1_Cylinder_Aabb() is not an option because it is for the shape Cylinder (which gridConnexion generalized without direct inheritance). >but this create a ScGridCoGeom6D instance. Is it equivalent to ScGeom Yes, ScGridCoGeom6D inherits from (i.e. "is a") ScGeom. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404520]: hollow glass sphere modeling using pfacet as a filler in cement paste
Question #404520 on Yade changed: https://answers.launchpad.net/yade/+question/404520 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, [1] is along the line I was mentioning, yes. Getting the total force on the indenter will just need to sum the forces on every elements of the pfacet surface. I have no experience in this sort of problems so I'm not the person to advise much more. Something I am wondering now is if the filler could not be simply a big sphere. Do you really need a hollow sphere? Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404550]: import particles to another script
Question #404550 on Yade changed: https://answers.launchpad.net/yade/+question/404550 Status: Open => Answered Bruno Chareyre proposed the following answer: Hello, the easiest solution is to not import. Why do you need to switch to another script instead of just one? I see plenty users doing that yet it is usually pointless (sometimes not). Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404534]: Control edge length in Tesselation Wrapper?
Question #404534 on Yade changed: https://answers.launchpad.net/yade/+question/404534 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi Luc, The cure to this disease is called "alpha shape", I have work in progress about that [1] but it is still in the state of independent pieces of code outside yade. It is a difficult problem for which and I don't see any simple workaround. Bruno [1] http://perso.3sr-grenoble.fr/users/bchareyre/nolist/nolist_temp/figTriax+Trou.png -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404513]: Is there any possibility that GPU be apllied in the calculation of YADE?
Question #404513 on Yade changed: https://answers.launchpad.net/yade/+question/404513 Bruno Chareyre proposed the following answer: > GPU implementation is almost impossible for YADE Not sure what it really means. If someone was going for GPU he would have to be prepared for big changes in the code of course, like for every other computer code. Wasn't it implicit in your question? >DEM is really fit for GPU implementation Not so sure. According to [1] GPU gives x7 speedup compared to single-core CPU. Which is pretty bad, we get nearly the same speedup from 12-cores OpenMP for large problems. In [2] the speedup was x10 but the comparison was unfair since the code tested on the GPU was a better re-implementation of the old sub-optimal code that was tested on the CPU, i.e. some small scale optimizations independent of CPU/GPU were also counted. Nothing really impressive overall. [1] https://arxiv.org/pdf/1301.1714 [2] https://hal.archives-ouvertes.fr/hal-00733674/document -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404513]: Is there any possibility that GPU be apllied in the calculation of YADE?
Question #404513 on Yade changed: https://answers.launchpad.net/yade/+question/404513 Status: Open => Answered Bruno Chareyre proposed the following answer: Hello, To the first question the answer is of course "yes". Unfortunately it is difficult (for me at least) to answer the second one before concrete implementation + benchmarking. Are you familiar with GPU implementations? Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404455]: damping in ViscElMat
Question #404455 on Yade changed: https://answers.launchpad.net/yade/+question/404455 Bruno Chareyre proposed the following answer: The porosity at equilibrium is not single valued, it is a range of possible porosities. Which one is reached along a particular loading path depends on every possible factor in principle, including numerical damping. As Jerôme suggests, you could introduce numerical daming only in the preparation phase if it helps reaching a particular porosity. Alternatively, for the same purpose, decrease the upper bound of wall velocity, wallMaxVel iirc (or the growth factor if internalCompaction=True); this will compact more gently then give looser states. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404363]: porosity calculation for clumps
Question #404363 on Yade changed: https://answers.launchpad.net/yade/+question/404363 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi Bettina, You find out that the volume(or mass) of the clump is calculated correctly (your third value). The porosity is not and that is expected, because the porosity function just take the volumes all clump members without looking at overlaps between them (you can even have negative porosity as a consequence). De facto, the documentation of this function is mentionning spheres and only that, not clumps. myClumpPorosity() is correct There is also voxelPorosity, which should be be correct at the price of extra computation time. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404328]: contact position at the end of the simulation
Question #404328 on Yade changed: https://answers.launchpad.net/yade/+question/404328 Bruno Chareyre proposed the following answer: Hi, Another quick comment. Using a PyRunner is a bit overkill for this particular need. If you want to run 5000 then export (Jan suggested that already but it you may miss it in his dense message), then the straight method is to really do it as it reads: O.run(5000,True) for i in O.interactions: ... yade.plot.addData(x1=x1,x2=x2,x3=x3) #see Jan's comment on not using addData plot.saveDataTxt(OUT1) -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404135]: intRadius(aabbenlargefactor) doesn't work
Question #404135 on Yade changed: https://answers.launchpad.net/yade/+question/404135 Bruno Chareyre posted a new comment: >"YADE graphical user interface offers a sub-step option (just check the corresponding box) that may allow you to check if the distant interaction exists at some point within a time step (before being possibly immediately erased by the contact law in the same time step" I don't think it is possible because everything happens in interactionLoop (one single sub-step). -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404135]: intRadius(aabbenlargefactor) doesn't work
Question #404135 on Yade changed: https://answers.launchpad.net/yade/+question/404135 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, > 1. Why is contact detection rely on mechanical behavior(In this case different normalCohesion value affect interaction number)? If cohesion is small the interactions are created then immediately broken for the tensile force exceeds the tensile strength. Thus one possible trick is to set high cohesion to create the ints, then set the relevant interaction->adhesion at the next step within a loop. We could add one option to initialize the force-free distance as the initial distance, it would make this particular situation a bit simpler (I think it is done this way in another (CPM?) contact model). > 2. In my understanding, setCohesionNow() merely set a maximum tensile strength between two particles(not a strength, but a maximum strength value), but in this occasion, after setting cohesion, some "strength" are generated between particles, why is that? To me "strength" means "maximum force". Not sure what "maximum strength" is. If you mean "some forces are generated", that is correct: the distant interactions are initially in traction (regardless of the strength except if it exceeds strength, in which case it is null). > 1. How can I answer to a specific person(or “floor” like #1,#2, etc), like Jan Stránský did in #3? "Reply to" the question from an email client, without erasing the replied-to (more a "lazy" than a "good" practice actually). > 2. Would it be possible to invite a specific person IIRC lanchpad lets you send 5 messages/day to other lauchpad accounts (click on the launchpad name -> contact this user). Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #404135]: intRadius(aabbenlargefactor) doesn't work
Question #404135 on Yade changed: https://answers.launchpad.net/yade/+question/404135 Bruno Chareyre posted a new comment: >Thus one possible trick is to set high cohesion to create the ints, then set the relevant interaction->adhesion at the next step within a loop. I forgot to add: make the initial configuration force-free if that's what you need, with i.phys.unp = i.geom.penetrationDepth. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] Yade installation questions
Dear David, Please use launchpad questions next time, instead of plain emails (*). On 11/15/2016 03:45 PM, Baker D.J. wrote: Gl1_Sphere attribute/command. Is that because I build Yade with GUI=OFF? Yes. "Gl" refers to openGL, hence graphical environment. It should be removed on users side. Obviously, manipulating 3D view is not an option in a script for cluster production. I then attempted to build the GUI version. The build (make) process completed successfully, however I found that “make install” failed looking for a file called “img_rc.py”. I cannot find any trace of this file in the Yale source distribution, and so I have no idea how to progress the GUI build. In this respect I have installed qt v4, however not PyQt4 – is the latter package required? I can't help for this one personally, sorry. Regards Bruno ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403713]: friction damping under different contact law
Question #403713 on Yade changed: https://answers.launchpad.net/yade/+question/403713 Bruno Chareyre proposed the following answer: Newton::damping should not be used at all if you are interested in granular flow (that is, not just quasistatic evolutions). hence no need to think about it. As for viscous damping not mimicing "some" physical behavior, it surely fails to describe an inelastic shock between brittle objects (for instance). The question here is not about viscous damping in itself but more about which material you are actually trying to simulate. It is usually a bit awkward to comment on decisions in DEM modeling without a prior knowledge of the real material properties. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403901]: Sample of pre-defined shaped particles
Question #403901 on Yade changed: https://answers.launchpad.net/yade/+question/403901 Bruno Chareyre proposed the following answer: Hi Bruna, Alternatively, it is possible is to place the particles one by one https://github.com/yade/trunk/blob/master/examples/triax-tutorial/script-session1.py#L67 to form clump templates, then they can be generated with random position/orientation using clumpCloud [1]. Not sure it is what you need though, to me the answer of Jan is also right. Bruno [1] https://github.com/yade/trunk/blob/master/examples/triax-tutorial /script-session1.py#L67 -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403764]: How to stop YADE from exiting automatically on PBS?
Question #403764 on Yade changed: https://answers.launchpad.net/yade/+question/403764 Bruno Chareyre proposed the following answer: Note that for the same reason your script is weird - at least unpredictable - independently of PBS: O.run() def addData(): [...] O.saveTmp() If you are lucky python will read the definition of addData() before the function is effectively used at first iteration, if you are not lucky you may get a "unknown function" error. As for the saveTmp(), nobody can tell if it will save the initial state or the state after 1-2 iterations. For sure, it is not saving the final state. B -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403764]: How to stop YADE from exiting automatically on PBS?
Question #403764 on Yade changed: https://answers.launchpad.net/yade/+question/403764 Bruno Chareyre proposed the following answer: As Jerome guessed, O.run() was the problem. "O.run()" means to python, almost, "spawn an independant process running DEM in the background and resume executing the script immediatly". At the end of the script the master process terminates. From a job scheduler point of view the job is finished, even if DEM is still running in the background. If you mean to wait until the DEM is finished the command is O.run(N,true) (the default is O.run(N,false)). Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403744]: Cohesive and viscous-elastic contact law
Question #403744 on Yade changed: https://answers.launchpad.net/yade/+question/403744 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, No, as far as I know. There is no law functor combining adhesion with this type of viscous damping. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403426]: how to compute the shear strain
Question #403426 on Yade changed: https://answers.launchpad.net/yade/+question/403426 Status: Open => Answered Bruno Chareyre proposed the following answer: Not sure about the sign problem you see. The 0.5 is because the normal contact spin is obtained by averaging bodies spin. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403469]: How to set the damping coefficient of a crowd of powders
Question #403469 on Yade changed: https://answers.launchpad.net/yade/+question/403469 Status: Open => Answered Bruno Chareyre proposed the following answer: >"your meaning that there is no physical damping" I meant that NewtonIntegrator::damping is a numerical trick to slow down the system. It does not correspond to any law of physics, so you better don't use that for realistic simulation of granular flow. > "kinetic energy of particles needs to be dissipated since the energy dissipation through friction damping during sliding may not be enough" This is a very common fairy tail, leading to some bad practice. I actually do undamped simulations quite often, in many cases it is completely safe (not always though), and sometimes computationally more efficient (because the loading rate can be increased). So, I'll not try to justify the above claim. Instead, I would say: a problem that "may" occur should not be fixed by making the numerical model physically inconsistent. I would start pushing the layer without damping. If you observe blatantly excessive bouncing of some elements then you can think of using viscous contacts, else you can just forget the issue I think. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403385]: Micro and Macro relation in CPM pararmeters
Question #403385 on Yade changed: https://answers.launchpad.net/yade/+question/403385 Bruno Chareyre proposed the following answer: Actually no, the micro-macro relations can always be put in a way such that they do NOT depend on particle sizes. In Yade, at least, it is the case (and it is obviously a good feature as you may found out). If you agree that cylinder size/aspect ratio should not play a role either, then interaction radius if any is the only remaining geometrical parameter, but it is in the end a contact parameter more than a geometrical parameter. What should bee kept the same in every geometry is porosity, coordination number, and fabric (an)isotropy. I think that's it. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403469]: How to set the damping coefficient of a crowd of powders
Question #403469 on Yade changed: https://answers.launchpad.net/yade/+question/403469 Bruno Chareyre posted a new comment: I meant to make damping zero after the initial deposit is stable. If yo umake it zero from the beginning the gravitational deposition will take ages. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403469]: How to set the damping coefficient of a crowd of powders
Question #403469 on Yade changed: https://answers.launchpad.net/yade/+question/403469 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, I would suggest to make Newton's damping 0 in such case, since it is not a physical damping. The question is if you need damping at all, I'm not sure of that. If you need some you could go for some sort of viscous contact damping. As for which contact viscosity would be best I'm afraid I have no clue. A parametric analysis compared with measurements could help to decide. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403426]: how to compute the shear strain
Question #403426 on Yade changed: https://answers.launchpad.net/yade/+question/403426 Bruno Chareyre posted a new comment: I forgot a multiplication by △t in the above equations. Let "v" be the rotation increment equal to spin times △t and the equations are correct. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403385]: Micro and Macro relation in CPM pararmeters
Question #403385 on Yade changed: https://answers.launchpad.net/yade/+question/403385 Bruno Chareyre proposed the following answer: Hi, Short answer: There is indeed no simple relationship between ks/kn and the material parameter called Poisson, but a dependency does exist (with both the solid phase Poisson coeff. and bulk scale Poisson) hence the conventional name. Longer answer: Since material parameters and interaction parameters in Yade are independent things (different functors will lead to different dependencies) it is not realistic to expect a physical consistency of the naming. In the end the parameter "Poisson" is just a value, it doesn't have to be ks/kn - only some particular functors may really use it to define ks/kn. Besides, I can't follow you on the question "Can we set a geometry which lead to macro parameters the same as micro ones?" If a particular micro-macro relationship (e.g. equality, in your case) exists the it should hold independently of the geometry of the specimen, else it is a physical aberration. So geometry is obviously not the right button to push to solve your problem. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403426]: how to compute the shear strain
Question #403426 on Yade changed: https://answers.launchpad.net/yade/+question/403426 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, I understand the meaning my not be immediate but it is actually relatively simple. Let w be the spin vector of the grain pair. Decompose this spin into normal and orthonormal components w= w_n + w_t. the rotation of any vector quantity attached to the pair can be approximated by v = v + w×v = v + w_n×v + w_t×v (cross products) In the doc the first cross-product is called (△uT) 2 and the second one is called (△uT) 1. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403204]: position of spheres in periodic cell
Question #403204 on Yade changed: https://answers.launchpad.net/yade/+question/403204 Status: Open => Answered Bruno Chareyre proposed the following answer: PBC are a bit complicated to explain. Please make sure to read the documentation on PBC and if something is still not clear be specific on what it is. The problem in you loop is the same as in for k in [1,2,3]: x=k the above loop is useless since it is equivalent to x=3 Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403166]: Code availability of Yuan2016, Sweijen2016
Question #403166 on Yade changed: https://answers.launchpad.net/yade/+question/403166 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, The code is available, yes. I think it will be available in a more user-friendly shape in less than a year when both devs will have their PhD defended (Chao Yuan defended very recently, Thomas Sweijen not yet). Cheers Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #403204]: position of spheres in periodic cell
Question #403204 on Yade changed: https://answers.launchpad.net/yade/+question/403204 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, 1) gives the absolute position as it is memory, it can be anywhere in unbounded infinite 3D space 2) gives for each particle its image particle in the reference period, so that all the wraped positions are close to each other (merely what you see in the 3D view) Note that your loops are odd, they result in position=O.bodies[-1].state.pos #without loop Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402989]: exporting the file of properties of spheres in simulation
Question #402989 on Yade changed: https://answers.launchpad.net/yade/+question/402989 Status: Open => Answered Bruno Chareyre proposed the following answer: Thanks for details Hien. > My other advice is to save anyway the full model (xml.gz) Sorry but... the "xml.gz" part is a very bad advice ;) "yade.gz" will reduce time and size big time. There is no reason to use xml. B -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402989]: exporting the file of properties of spheres in simulation
Question #402989 on Yade changed: https://answers.launchpad.net/yade/+question/402989 Status: Open => Answered Bruno Chareyre proposed the following answer: There are plenty ways to do that, yes. I suggest to first clarify which data you need, and for instance to write a loop to print it in the terminal: for b in O.bodies: print b.state.pos[0], b.state.pos[1], b.state.pos[2], b.material.friction, etc. When it has the shape you want you do exactly the same but writing to a file. I hope it helps. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402893]: Ids order in O.interactions
Question #402893 on Yade changed: https://answers.launchpad.net/yade/+question/402893 Bruno Chareyre proposed the following answer: The force is always applied by i.id1 on i.id2. No need to use the normal. Does it answer your question? -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402893]: Ids order in O.interactions
Question #402893 on Yade changed: https://answers.launchpad.net/yade/+question/402893 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi Nejib, there is no convention. The fact that O.interactions[x,y] is the same as O.interactions[y,x] is a desirable feature, isn't it? Both ways give the same interaction, hence the same particular ordering of ids. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402828]: Control the normal stress fluctuation in the direct shear box modeling
Question #402828 on Yade changed: https://answers.launchpad.net/yade/+question/402828 Status: Open => Answered Bruno Chareyre proposed the following answer: One possible strategy is to simply impose a force on the top plate and let it equilibrate (it needs to have a mass). Is is simple but it does not offer the best control usually, due to the inertia of the plate. Another strategy is to impose de velocity (or displacement): v = f(sm-st), with "sm" the measured force and "st" the target force. The remaining question is how to define f. One way is to calculate the cumulated stiffness K of all plate-particle interactions, and to use it in: v=(sm-st)/(K*O.dt) which will produce a nearly constant force. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402752]: Geometry law between a sphere and a wall for the cohesive frictional law
Question #402752 on Yade changed: https://answers.launchpad.net/yade/+question/402752 Status: Open => Answered Bruno Chareyre proposed the following answer: Ok. So, first of all, I think Walls are not very commonly used and not all families of functors may handle them. namely, ScGeom6D does not support walls, as you certainly found out. The more realistic choice is between Facet and Box (a Box with 0-thickness in one direction is a plane/wall, right? ), and they should give exactly the same result in principle. Hence the choice: Ig2_Box_Sphere_ScGeom6D or Ig2_Facet_Sphere_ScGeom6D I personally implemented the "6D" functors hence I can tell the Box variant was more seriously tested (by me), but hopefully both ways are ok. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402752]: Geometry law between a sphere and a wall for the cohesive frictional law
Question #402752 on Yade changed: https://answers.launchpad.net/yade/+question/402752 Status: Open => Needs information Bruno Chareyre requested more information: Hi, Your question is about walls but your script introduces facets. Could you clarify the question? -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402655]: How many digits is the significant figure of the position?
Question #402655 on Yade changed: https://answers.launchpad.net/yade/+question/402655 Status: Open => Answered Bruno Chareyre proposed the following answer: Approximately 16 (more precisely 15.95) digits for double precision format. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #402426]: fill a box with polyhedra with size distribution (psdSizes, psdCumm )
Question #402426 on Yade changed: https://answers.launchpad.net/yade/+question/402426 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, If you mean a makeCloud function specialized for polyhedra I would say no. But actually you can use makeCloud to generate a sphere cloud and use this list to place polyhedra of the same size (say, each polyhedron is represented by the smallest sphere enclosing it). list=makeCloud() for s in list: O.bodies.append([convert sphere to polyhedron here]) Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #372306]: Stability of PFV
Question #372306 on Yade changed: https://answers.launchpad.net/yade/+question/372306 Bruno Chareyre posted a new comment: Hi Benoit, The stability problem is explained in E. Catalano PhD, where an expression of the stability condition is derived. The time step depends linearly (iirc) on viscosity (inverse), and mass. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #384899]: VoxelPorosity
Question #384899 on Yade changed: https://answers.launchpad.net/yade/+question/384899 Status: Open => Answered Bruno Chareyre proposed the following answer: Did you make basic checks like: how many spheres remain after removal? where are they located? what is their radius? Most likely the problem is in your code, not in regularHexa() nor voxelPorosity(). Bruno p.s. Note that you exemple is not self-contained, hence nearly useless. rmin, rmax are not defined. See https://yade-dem.org/wiki/Howtoask -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #239596]: FlowEngine is not defined in yade-daily - On Ubuntu 13.10
Question #239596 on Yade changed: https://answers.launchpad.net/yade/+question/239596 Bruno Chareyre posted a new comment: Hi Son, Modifying the source code to escape a compile error is the best starting point for big troubles. If you have compile errors please fill a bug report (this problem you have is now solved in latest trunk IIRC), and please don't append more questions to existing ones. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #314841]: Cannot figure out how to use 2D TriaxialStressController
Question #314841 on Yade changed: https://answers.launchpad.net/yade/+question/314841 Bruno Chareyre proposed the following answer: Hi, Imposing zero stress can be a problem when the sample-boundary contacts have no tensile strength (the typical case). This is because no contact at all satisfy the zero-stress condition, then the wall can be just floating around, in turns giving a wrong cross-sectional area, and then a wrong stress evaluation. I suggest to impose a small non-zero value. Would it change the problem? Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #313437]: The relationship between stress tensor and stress vector
Question #313437 on Yade changed: https://answers.launchpad.net/yade/+question/313437 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, getStress gives an ordinary 3x3 tensor, not 6x6. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #309164]: Plan with no interactions
Question #309164 on Yade changed: https://answers.launchpad.net/yade/+question/309164 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi Bruna, You could always add a plane, get force, then reload the state before adding the plane and continue. But overall it doesn't make much sense. The actual interactions in a sheared layer are unrelated to what you would get by forcing a plane in between (leading to huge forces since the plane will cut spheres randomly). It needs more details to understand what you really have in mind. In fact there is non discontinuity of stress in a shear band, so you can as well use the stress exerted on the top boundary, it has to be the same. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #298001]: Input Kn and Ks directly
Question #298001 on Yade changed: https://answers.launchpad.net/yade/+question/298001 Bruno Chareyre posted a new comment: > Is there any paper/thesis/publication that discusses this advantage of YADE's formulation? It is mentioned in many of our papers but it is never the main topic; I don't know any third party paper that would discuss this specific point in detail either (OTOH I'm not sure it is worth a paper...). If you just mean a paper where a few sentences about kn,ks normalization are included it can be found. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295810]: CPM contact parameters (stiffness & friction angle)
Question #295810 on Yade changed: https://answers.launchpad.net/yade/+question/295810 Status: Open => Answered Bruno Chareyre proposed the following answer: Hello, MatchMakers are definitely the answer to your question. It is not possible to hardcode in the source code, in advance, every possible relationships. Changing manually the contact parameters is ok. The problem is it is difficult to control what happens to newly created contacts with this method. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295814]: MatchTracker for Cohesion
Question #295814 on Yade changed: https://answers.launchpad.net/yade/+question/295814 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi Seti, I suggest having a look at those guidelines, especially points 2,3 and 6: https://yade-dem.org/wiki/Howtoask Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295745]: Do bending or twisting moment contribute to cohesive bond breakage ?
Question #295745 on Yade changed: https://answers.launchpad.net/yade/+question/295745 Bruno Chareyre proposed the following answer: Yes. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295745]: Do bending or twisting moment contribute to cohesive bond breakage ?
Question #295745 on Yade changed: https://answers.launchpad.net/yade/+question/295745 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, The contact moments can have upper bounds (strength) but for the moment the upper bound can only take the form Mmax=a*Fn where "a" is user defined coefficient. Also rotational plasticity will not interfere (will not "break") in the calculation normal/shear terms, as you found out. It is not very difficult to generalize the current code if someone has a precise idea in mind. >when the contact is in tension, i.e. Fn < 0 , does the equation above actually reduce the value of maxFs ? Yes indeed, if cohesion and friction are combined this is what happens. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295685]: the difference between ScGeom and L3Geom
Question #295685 on Yade changed: https://answers.launchpad.net/yade/+question/295685 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, L3Geom is unmaintained and it has known bugs. Better don't use it. That is the first and main difference. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295666]: dense sample/triaxial test
Question #295666 on Yade changed: https://answers.launchpad.net/yade/+question/295666 Status: Open => Answered Bruno Chareyre proposed the following answer: I suggest a look at the triax tutorial. It is doing what you ask in a better way (the exact density can be prescribed). Bruno https://github.com/yade/trunk/blob/master/examples/triax-tutorial /script-session1.py -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295587]: Box Free Oscillation
Question #295587 on Yade changed: https://answers.launchpad.net/yade/+question/295587 Bruno Chareyre proposed the following answer: Yes, in fact the "Box" class is practically handling thick planes, not boxes, in the sense that they can't contain anything. Sphere inside box is an algorithmic nightmare in Yade since it would need multiple contact interactions between the two same bodies. As suggested by Jan a practical approach is simply the one used in Triaxial tutorials! 6 Box's to bound one box. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295592]: Problems for allowBiggerThanPeriod
Question #295592 on Yade changed: https://answers.launchpad.net/yade/+question/295592 Bruno Chareyre proposed the following answer: > is there any possible for applying large deformation? I mean shear a sample without ceasing. I assume you refer to true 3-periodic boundary conditions. Then the answer is: - In theory yes, it is possible - In Yade no, it is not possible although we are not very far from it and incomplete steps in this direction have been taken in the past - In other DEM codes, usually not. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295521]: contact-normal in Yade
Question #295521 on Yade changed: https://answers.launchpad.net/yade/+question/295521 Status: Open => Answered Bruno Chareyre proposed the following answer: >I am afraid that this method may get fabric tensor calculation into trouble. Fabric tensor is based on squared normal, so the problem disappears: n and -n gives the same fabric. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295592]: Problems for allowBiggerThanPeriod
Question #295592 on Yade changed: https://answers.launchpad.net/yade/+question/295592 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, > allowBiggerThanPeriod can be used for large deformation Absolutely not, as you find. This is for large bodies, not large deformations. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295520]: Exception occured: std::bad_alloc
Question #295520 on Yade changed: https://answers.launchpad.net/yade/+question/295520 Bruno Chareyre proposed the following answer: There should not be any memory cost for this function. Why is there an indentation before "O" in: ### add top box O.bodies.append(utils.box(center=(xinf-thickness+Rmean/10,3*(ysup/4)+2*Rmean,zsup/2) It could be a problem. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295218]: Can't load the Launchpad of Yade
Question #295218 on Yade changed: https://answers.launchpad.net/yade/+question/295218 Bruno Chareyre proposed the following answer: It was down for a few hours indeed, then fixed before Jérôme tested it. B -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295158]: Can we create a rigid body using only particles, rather than importing a gts surface.
Question #295158 on Yade changed: https://answers.launchpad.net/yade/+question/295158 Bruno Chareyre proposed the following answer: > A ball made up of spheres, bouncing in a box, which is also made of spheres. Bullet made up of spheres hitting a concrete wall composed of spheres. Yes of course, you can dot that. Why not? Could you clarify your question? Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295035]: Use of i.phys.unp to get null forces initially
Question #295035 on Yade changed: https://answers.launchpad.net/yade/+question/295035 Bruno Chareyre proposed the following answer: >So should unp be changed after few iterations, to get more accurate normal contact force Certainly not. Why would it be "more accurate"?! B -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #295035]: Use of i.phys.unp to get null forces initially
Question #295035 on Yade changed: https://answers.launchpad.net/yade/+question/295035 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, I just updated the doc but it is not visible in the online pages yet. The normal force in the first equation of the documentation [1] will look like this: fn=max(kn*(un-unp),an) Bruno [1] https://yade-dem.org/doc/yade.wrapper.html#yade.wrapper.Law2_ScGeom6D_CohFrictPhys_CohesionMoment -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #294519]: Forcing InsertionSortCollider execution
Question #294519 on Yade changed: https://answers.launchpad.net/yade/+question/294519 Bruno Chareyre posted a new comment: >it is not possible to actually update verletDist during the simulation It is possible of course. The main problem is that the collider is smart: if the grains do not move (as in your exemple) there is no need to detect new collisions. Consequence: the lines after "# Attempting to introduce distant interactions" are not doing anything You can check in the 3D view, the bounds are not modified. The new sizes would be accounted for only after the next colliding step (give a velocity to one body, bounds will change). >As for runEngine(), I confirm adding O.runEngine(O.engines[1]) before O.step() is enough I see. runEngine() skips the check InsertionSortCollider::isActivated(), so you get something even if it would not normally run. In the long run, I think this question of creating distant interactions should be solved without the collider because it has different requirements compared to detecting collisions. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #294958]: meniscus fusion
Question #294958 on Yade changed: https://answers.launchpad.net/yade/+question/294958 Status: Open => Answered Bruno Chareyre proposed the following answer: Yes you are right. However, the maximum saturation is more around 5% if I remember correctly. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #294845]: Multicore Yade
Question #294845 on Yade changed: https://answers.launchpad.net/yade/+question/294845 Bruno Chareyre proposed the following answer: Hi Pavel, Yade has profiling features which should be useful in your case. Please see https://yade-dem.org/doc/prog.html#timing Until now I would follow Anton. If you did not implement parallel versions of the following functions then of course giving more cores is not magically parallelizing them : PyRunner(command='SaveData()',virtPeriod=1./1000), ForceResetter(), VppNeighbours(), VppPressureForce(), PyRunner(command='ChangeMask()',iterPeriod=1), PyRunner(command='ChangePos()',iterPeriod=1), PyRunner(command='ChangePosFoil()',iterPeriod=1), PyRunner(command='PreChangePos()',iterPeriod=1), PyRunner(command='PressureOn()',iterPeriod=1), PyRunner(command='PressureTimeOn()',iterPeriod=1), -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #293955]: How to generate dense packing of aggregates
Question #293955 on Yade changed: https://answers.launchpad.net/yade/+question/293955 Bruno Chareyre posted a new comment: Please, one question at a time. If this one is answered mark it closed. If you have other questions open them as new questions. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #294529]: random orientation of chained cylinders
Question #294529 on Yade changed: https://answers.launchpad.net/yade/+question/294529 Bruno Chareyre requested more information: Hi, The initial properties of your fiber distribution should be defined more precisely before any concrete implementation can be conceived, seems to me. For instance, Klaus talks about generating random points then connecting them. A valid answer in view of the current question, yet maybe not consistent with what your real needs are (lengths, initial curvature,...). Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #294519]: Forcing InsertionSortCollider execution
Question #294519 on Yade changed: https://answers.launchpad.net/yade/+question/294519 Bruno Chareyre posted a new comment: Hi Jérôme, Setting verletDist=0 will make the collider run at each iteration. Does it really do something with runEngine? I would expect the collider to just return if not a collliding situation. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #293758]: handling post-breakage rolling in a cohesive contact
Question #293758 on Yade changed: https://answers.launchpad.net/yade/+question/293758 Bruno Chareyre posted a new comment: "If you don't understand" -> "I don't understand" A weird auto-completion got me there. -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #293758]: handling post-breakage rolling in a cohesive contact
Question #293758 on Yade changed: https://answers.launchpad.net/yade/+question/293758 Status: Open => Answered Bruno Chareyre proposed the following answer: Hi, > if the shearForce.norm() > shearAdhesion, break the cohesion and now as the friction angle is bigger, the movement is more difficult. If you don't understand why friction angle should be bigger. Also, what is the relation to rolling. It is not very clear to me ,sorry. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #293445]: Polyhedra Elements, Infinite Loop.
Question #293445 on Yade changed: https://answers.launchpad.net/yade/+question/293445 Status: Open => Needs information Bruno Chareyre requested more information: Hi, Thanks for the script but in fact it does not show the bug, since you add "After calling O.step() for several times, this bug will occur." It may sound like a dumb question but can you really try to make the script hang by itself? The reason could be there. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #293295]: Want to connect all the particles in an aggregate using cohesive bond
Question #293295 on Yade changed: https://answers.launchpad.net/yade/+question/293295 Bruno Chareyre proposed the following answer: Changing the adhesion of existing interactions is not a problem at all. After the first iteration: i.phys.normalAdhesion = something i.phys.shearAdhesion = another thing i.phys.unp = 0 #to make it initially force free, as in the link quoted in #3 The only problem was that you were changing material parameters instead of changing interaction parameters (as found by whr, even if he suggests a different workaround). -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #293096]: incrememntal compression with engine
Question #293096 on Yade changed: https://answers.launchpad.net/yade/+question/293096 Status: Open => Answered Bruno Chareyre proposed the following answer: If I understand correctly you suggest that this loop is not doing anything: for i in range(0,int(-1e-6*stress_max)): if ( abs(triax.goal2) < abs(stress_max) ): O.run(200,True) triax.goal2+=delta_stress I would thus suggest to double-check what it does. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #292663]: capillary pressure
Question #292663 on Yade changed: https://answers.launchpad.net/yade/+question/292663 Bruno Chareyre posted a new comment: Thanks Christian, useful answer for me to. We definitely need to consolidate this interpolation part and make it more self contained (dropping the data files download and let people generate them with a procedure we provide). Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #290950]: Particles crossing the wall
Question #290950 on Yade changed: https://answers.launchpad.net/yade/+question/290950 Bruno Chareyre posted a new comment: @Jérôme: the workaround is to not have interactions with the large body at step 0 -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #290950]: Particles crossing the wall
Question #290950 on Yade changed: https://answers.launchpad.net/yade/+question/290950 Status: Open => Answered Bruno Chareyre proposed the following answer: You found a known bug for large objects in periodic simulations. The bug seems to show up especially when some interactions exist at iteration 0, those ones are not detected correctly (this is your case, where you have large overlaps in the initial geometry). I find that the problem disappear in your case if line 149 is changed as follows: top=box( center=(length/2.0,yExt - yLim + thickness*2,width/2.0) ,extents=(4*length,thickness/2.0,width) ,wire=False ,material='plate_mat' ) Besides, there is absolutely no need to define a thickness for your plates. Thickness zero is ok (mind that the default will be zero though). Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #291538]: triaxial test:Law2_ScGeom_FrictPhys_CundallStrack
Question #291538 on Yade changed: https://answers.launchpad.net/yade/+question/291538 Bruno Chareyre posted a new comment: Please don't ask the same question multiple times (actually there is not even a question in this one ;) ). If you don't receive answers you may check (especially point 3): https://yade-dem.org/wiki/Howtoask Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #290070]: TesselationWrapper and porosity
Question #290070 on Yade changed: https://answers.launchpad.net/yade/+question/290070 Status: Open => Answered Bruno Chareyre proposed the following answer: b.volume=TW.volume(b.id) is not the same in terms of scope, and the problems you have are mainly scope problems. >"The question is how we can export a user defined parameter the same way as >the strain?" We cannot, since there cannot be a specific export function implemented in advance for every possible _user defined_ parameter. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #290070]: TesselationWrapper and porosity
Question #290070 on Yade changed: https://answers.launchpad.net/yade/+question/290070 Bruno Chareyre proposed the following answer: It must be a scope issue, try something around globals()["TW"].volume(id)? Else it should be possible to add the attribute: b.volume=TW.volume(b.id) #then use it to export -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #290070]: TesselationWrapper and porosity
Question #290070 on Yade changed: https://answers.launchpad.net/yade/+question/290070 Bruno Chareyre proposed the following answer: I am assuming that you know how to export any value attached to particles using VTK exporter (else please see the doc, this one is well documented). Then the only question is how to compute porosity for one particle. I don't understand the equation (1.13097e-4)/TW.volume(i.id) If Vi the volume of the sphere "i", then I would think the porosity is [1 - Vi/TW.volume(i.id)] Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #289169]: How can I get local fluxes in flowEngine ?
Question #289169 on Yade changed: https://answers.launchpad.net/yade/+question/289169 Status: Open => Answered Bruno Chareyre proposed the following answer: There is a c++ function which assign the average velocity of each cell. FlowBoundingSphere.ipp:96 void FlowBoundingSphere::averageRelativeCellVelocity() It has no python wrapping though, but maybe you can manage to do that? (I'll not be able to help on that at the moment) Note that velocity is not a variable of the problem, only fluxes are really computed. This velocity is a post-processed quantity. Also note that it is the velocity relative to the solid phase, so that it is usually weird (since one intuitively expect to see absolute velocities in a fixed reference frame). Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #288586]: Can Yade run on GPU?
Question #288586 on Yade changed: https://answers.launchpad.net/yade/+question/288586 Status: Open => Answered Bruno Chareyre proposed the following answer: Hello, No, there is no GPU support. Bruno -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
Re: [Yade-users] [Question #288118]: What does the list returned by getConstrictionsFull() stands for in FlowEngine() ?
Question #288118 on Yade changed: https://answers.launchpad.net/yade/+question/288118 Bruno Chareyre proposed the following answer: My mistake, sorry. You can now update: cd trunk git pull cd [buildDir] make install -- You received this question notification because your team yade-users is an answer contact for Yade. ___ Mailing list: https://launchpad.net/~yade-users Post to : yade-users@lists.launchpad.net Unsubscribe : https://launchpad.net/~yade-users More help : https://help.launchpad.net/ListHelp
