coefficient of friction for wall

Submitted by adabubu on Mon, 12/06/2010 - 16:23

Hi, Christoph,
I read some papers about Janssen's equation. I have noticed that the friction angle of wall is involved in this equation. i know there is a command to define coefficient of friction, e.g. ' fix all property/global coefficientFriction peratomtypepair 1 0.7 ' , but as i understand, this 'coefficientFriction' should be for particles, so, how could i define coefficient of friction for wall?

thanks,

adabubu

Forums:

ckloss's picture

ckloss | Mon, 12/06/2010 - 16:42

Hi adabubu,

to my knowledge there is no generally accepted relationship between microscopic parameters (which are needed for the simulation input) and macroscopic parameters such as angle of wall friction. What people typically do is to make an educated guess for the microscopic parameters and calibrate the parameters to match the measured macroscopic property, such as angle of wall friction

Christoph

raguelmoon's picture

raguelmoon | Sun, 12/12/2010 - 14:22

Hi,
Is it right way of setting particle-wall friction zero and particle-particle friction 0.1?

fix all property/global coefficientFriction peratomtypepair 2 0.1 0.0 0.0 0.0
Then i apply wall-type 2.
Thanks in advance,
Ram

Ram

raguelmoon's picture

raguelmoon | Mon, 12/13/2010 - 16:13

Hi H,
You can apply wall friction to :
fix xwalls all wall/gran/hertz/history 1 0 xplane 0 0.225 2 (where 2 is wall-type 2)
or to mesh wall :
fix cad all mesh/gran wall.stl 2 0.01 0.0 0.0 0.0 0. 0. 0.( where 2 is wall-type 2)

Cheers,
Ram

Ram

adabubu | Thu, 12/16/2010 - 15:57

Thank you. it works.
but i have a following question. As i understand, the coefficient of friction (wall-particle) derived from the numerical ananlysis (shear stress divided by normal stress) should be equal to it defined in the input file, but i could not get it. So my question is what kind relastionship between these two kind of friction (from definition and numerical results) is. Thanks,
ada

ckloss's picture

ckloss | Fri, 12/17/2010 - 12:10

I don't get the question... What do you mean by friction friction from definition? As I said, there is no generally accepted relationship between microscopic parameters (which are needed for the simulation input) and macroscopic parameters such as angle of wall friction

Christoph

adabubu | Fri, 12/17/2010 - 14:12

hi, Christoph, sorry i didn't explain very clearly. I did a simulation of pouring problem and in my input file there are two coefficients of friction, one is for particle-particle, another is for wall-particle. After the simulation, i get the stress distribution of wall, then i calculate a coefficient of friction for wall from normal stress and shear stress of wall. If i understand correctly, this calculated coefficient of friction is the coefficient of friction between the wall and particles, then it should be equal to the coefficient of friction for particle-wall in the input file; but they are not same, so i want to konw what the problem is in my simulation.

Thank you

Ada

ckloss's picture

ckloss | Fri, 12/17/2010 - 15:05

>>but they are not same
What you are doing is parameter identification. As I said, there is no direct link between micro-and macro properties!

Christoph

alexander.polson | Fri, 01/14/2011 - 23:17

Hi,

I am a bit late for this discussion, but I would also be interested to be able to specify a coefficient of friction for the wall (mu_wall) which is different from mu_particle in LIGGGHTS. Did I understand correctly from the above threads that there is not a way to do this in LIGGGHTS?

If it cant be done in LIGGGHTS, a case for this might be where the particle and wall materials are different, e.g. iron ore which abrades against a smooth/slippery polymer liner. In e.g. PFC3D, you can do this, and then the effective contact mu becomes min(mu_wall,mu_particle).

If there is a way to do this, could someone please tell me the command?

Thank you

Alexander

ckloss's picture

ckloss | Sat, 01/15/2011 - 03:18

yes, you can do that, the coefficient of friction is specified per atom type pair, i.e. wall is type 1 and particle is type 2, you specify µ11, µ22 and µ12=µ21

Look up the doc for pair_gran, or see this thread: node/134

Cheers,
Christoph

diego.peinado's picture

diego.peinado | Sat, 01/15/2011 - 20:42

Hello Ada:
The Coulomb attrition condition states that the tangential stress in static conditions is equal or less than the product of the friction coefficient times the normal stress. If you define the coefficient of static friction (COF) in your script file, and then calculate a mean or averaged COF from output results, the latter should have a value smaller than the former.
Please, can you give us this result? (I mean if the calculated COF is less than the defined COF)
Thanks,

Diego

htm | Mon, 12/13/2010 - 14:48

Hi Ram
can you extend on what you mean by applying wall type-2? Where on the documentation do you find this? Could you post the lines of your code showing how you set particle-wall friction to zero?
Thanks a lot
H