ESyS-Particle

Experiences with ESYS and Damping model question

Asked by Diego Peinado Martín

First of all, I have to say that thanks to Dion, and all the development team of ESYS. It is a very complete and easy software package for DEM simulation. I've just finished a simulation consisting in a 2D box with a vibrating bottom wall. It has 380 small particles (1 mm) and 1 big particle (3 mm). The vibration is sinusoidal x=a.sin(w.t) with a = 3 mm and w = 100 1/seg, so the acceleration over gravity is Gamma = a.w² / g = 3.06. The time to setup the problem and of course to learn how to do it has been very short. The segregation caused by the vibration can be seen on the model. Thanks again for setting this package as open source.

After that I have one question. I've been reading some part of code (the files of the Model directory). As far as I've understood, the forces between particles are calculated in: CElasticInteraction, CFrictionInteraction, CRotElasticInteraction, CRotFrictionInteraction, ... and the corresponding for bonded interactions.

Here is not included the damping. There are another two classes CDamping and CRotDamping, that calculate the damping forces. If it is required to have linear damping as well as rotational damping then two different instances should be made, one for each type of damping.

My doubt is the following. As I can see, the damping is calculated for each particle, and not for a contacting pair of particles. It is much as a general damping for abolute velocities. Well, there is a field that has to do with a reference velocity that I did not see how it's used. The question is: the calculated damping uses the absolutes velocities and therefore is a general viscosity, or there is in other classes a mechanism, so it is used this reference velocity between particles so it is implemented as a contact viscosity.

Thanks in advance, and best regards

Question information

Language:
English Edit question
Status:
Solved
For:
ESyS-Particle Edit question
Assignee:
No assignee Edit question
Solved by:
Dion Weatherley
Solved:
Last query:
Last reply:
Revision history for this message
Best Dion Weatherley (d-weatherley) said : 		#1

Hi Diego,

Thanks for the kind words and endorsement of ESyS-Particle. I'm sure I speak on behalf of the whole development team when I say that it is very gratifying to hear that others appreciate the software and find it useful for their research.

Regarding your question, you are quite correct that the LinDamping and RotDamping interactions are "body forces" that act on particles individually. For linear damping, the damping force acting on particle i is:

F_i = -1 * visc * V_i * m_i

where visc is the coefficient of viscosity (supplied as an argument to LinDampingPrms), V_i is the current velocity of the particle and m_i is the mass of the particle. A similar damping moment calculation (proportional to the angular velocity of the particle) is applied when using RotDamping. The reference velocity (vref) is currently not used and its value is set equal to zero. It is an artifact of an experimental damping method dating back some years.

Whilst this so-called "artificial viscosity" is quite useful to damp kinetic energy in the model it does not achieve the same type of damping as a spring-dashpot interaction acting on bonded particle-pairs. A spring-dashpot will damp small oscillations (often numerical in origin) between two bonded particles. Unfortunately ESyS-Particle currently does not implement spring-dashpot interactions of any form (although it is very high on the list of new interaction types!).

I hope this helps and have fun with ESyS-Particle!

Cheers,

Dion.

Revision history for this message
Diego Peinado Martín (diego-peinado) said : 		#2

Thanks Dion Weatherley, that solved my question.