Question: Normal force and tangential force in EDEM calculation.

Yu Mu

Hello everyone,

I have two questions about contact force：

1.I see in the help document this sentence：

I wonder if this means that when I use the H-M model, the normal force derived from edem does not include damping force? Is the tangential force the same? Or does the derived normal force and tangential force include spring force and damping force?

2.Does the tangential damping force participate in the torque calculation of particle rotation?

Thank you in advance for your answers!

Zilong Qiu

Yu Mu said:

Hello, thank you for your answer！ Then, what I want to know is that when I export "normal force" or "tangential force" in the post-processing part (analyst), does the force given by edem include the resultant force of elasticity and damping?

hello Yu,

 

If I am right, only the undamped part is included when you export the normal force from the EDEM analyst (contact). But the damped part is considered in the particle motion calculation.

 

Regards,

Zilong

jpmorr

I'm not sure that statement is correct as EDEM returns the damped contact forces. Here's a simple test of a particles impacting a rigid elastic wall - one with damping and one without. If this statement was correct, the normal force plotted on each of these curves would be the same, but it's clear the damped normal force is being plotted.

 

I think this needs correction or maybe clarification if it is referring to something else, but from what I can tell, whether it's via the analyst, csv export or EDEMpy, that it's the damped force that is available.

I have not tested tangential forces, but i don't imagine that they are any different.

 

JP

 

Stephen Cole

Hi Yu, JP

EDEM calculates the damped force (damped force = undamped force + damping component) and uses this to determine the motion of the particles (e.g. effect of restitution value).  

However only the undamped force is reported in the EDEM Analyst and when transferring the forces to MBD solvers.

The reason for this is the real physical force at the contact is the spring force, the damping is an approximation in DEM so the convention is that the 'real' force is reported.

Including damping (restitution value) influences the motion which influences the force however documentation refers to how this force is then reported rather than how it is calculated.

 

For example in the API we have the following where F_n is the Normal force and F_nd is the damping force.  

 

You could return 0 for calculatedUnsymNormalForce and then the calculated and reported forces would be identical without any change in material behaviour.

 

Regards

Stephen

 

 

Yu Mu

Stephen Cole_21117 said:

Hi Yu, JP

EDEM calculates the damped force (damped force = undamped force + damping component) and uses this to determine the motion of the particles (e.g. effect of restitution value).  

However only the undamped force is reported in the EDEM Analyst and when transferring the forces to MBD solvers.

The reason for this is the real physical force at the contact is the spring force, the damping is an approximation in DEM so the convention is that the 'real' force is reported.

Including damping (restitution value) influences the motion which influences the force however documentation refers to how this force is then reported rather than how it is calculated.

 

For example in the API we have the following where F_n is the Normal force and F_nd is the damping force.  

 

You could return 0 for calculatedUnsymNormalForce and then the calculated and reported forces would be identical without any change in material behaviour.

 

Regards

Stephen

 

 

Hello, thank you for your answer！ Then, what I want to know is that when I export "normal force" or "tangential force" in the post-processing part (analyst), does the force given by edem include the resultant force of elasticity and damping?

Zilong Qiu

Yu Mu said:

Hello, thank you for your answer！ Then, what I want to know is that when I export "normal force" or "tangential force" in the post-processing part (analyst), does the force given by edem include the resultant force of elasticity and damping?

hello Yu,

 

If I am right, only the undamped part is included when you export the normal force from the EDEM analyst (contact). But the damped part is considered in the particle motion calculation.

 

Regards,

Zilong

jpmorr

Stephen Cole_21117 said:

Hi Yu, JP

EDEM calculates the damped force (damped force = undamped force + damping component) and uses this to determine the motion of the particles (e.g. effect of restitution value).  

However only the undamped force is reported in the EDEM Analyst and when transferring the forces to MBD solvers.

The reason for this is the real physical force at the contact is the spring force, the damping is an approximation in DEM so the convention is that the 'real' force is reported.

Including damping (restitution value) influences the motion which influences the force however documentation refers to how this force is then reported rather than how it is calculated.

 

For example in the API we have the following where F_n is the Normal force and F_nd is the damping force.  

 

You could return 0 for calculatedUnsymNormalForce and then the calculated and reported forces would be identical without any change in material behaviour.

 

Regards

Stephen

 

 

Hi Stephen,

From the two screenshots I attached, you can see that EDEM Analyst and other methods of exporting results export the damped force - the damping has changed the magnitude and shape of the normal force curve. If it was reporting the undamped force, then both curves would be identical.

While EDEM is internally differentiating damped and undamped as you show with the API (and is described in the API documentation I think), the exported results simply provide the damped force.

This confusion is coming from the documentation which does not state what the attribute described in the appendix is referring to. It does not make clear if this contact normal force description is how EDEM deals with the normal force internally or if this relates to what result the user will be provided. There is no option when exporting results data to access the undamped force and the damping term separately, so exported data is always damped?

If I'm reading this documentation (and not the API documentation) I'm going to assume that this is the description of the contact result that I plot and export.

I'm not sure what context the original post was referring to, whether it was internal usage or exported results, but it appears the two use cases need to be more clearly described in the documentation to avoid further confusion.

Regards,

JP

Yu Mu

Stephen Cole_21117 said:

Hi Yu, JP

EDEM calculates the damped force (damped force = undamped force + damping component) and uses this to determine the motion of the particles (e.g. effect of restitution value).  

However only the undamped force is reported in the EDEM Analyst and when transferring the forces to MBD solvers.

The reason for this is the real physical force at the contact is the spring force, the damping is an approximation in DEM so the convention is that the 'real' force is reported.

Including damping (restitution value) influences the motion which influences the force however documentation refers to how this force is then reported rather than how it is calculated.

 

For example in the API we have the following where F_n is the Normal force and F_nd is the damping force.  

 

You could return 0 for calculatedUnsymNormalForce and then the calculated and reported forces would be identical without any change in material behaviour.

 

Regards

Stephen

 

 

Hello, I still have a question about this: when sliding occurs, the tangential force calculated by edem is sliding friction. Will this force be reported? Or is it not reported as damping force?

Stephen Cole

J.P. Morrissey said:

Hi Stephen,

From the two screenshots I attached, you can see that EDEM Analyst and other methods of exporting results export the damped force - the damping has changed the magnitude and shape of the normal force curve. If it was reporting the undamped force, then both curves would be identical.

While EDEM is internally differentiating damped and undamped as you show with the API (and is described in the API documentation I think), the exported results simply provide the damped force.

This confusion is coming from the documentation which does not state what the attribute described in the appendix is referring to. It does not make clear if this contact normal force description is how EDEM deals with the normal force internally or if this relates to what result the user will be provided. There is no option when exporting results data to access the undamped force and the damping term separately, so exported data is always damped?

If I'm reading this documentation (and not the API documentation) I'm going to assume that this is the description of the contact result that I plot and export.

I'm not sure what context the original post was referring to, whether it was internal usage or exported results, but it appears the two use cases need to be more clearly described in the documentation to avoid further confusion.

Regards,

JP

Hi JP,

Could you let me know which physics model and method of including/excluding the damping?

For clarification I would expect that including damping to change the shape of the curve as it influences the motion.  However the undamped force is the one reported, this is all of the Analyst (visualising the contacts, plotting graphs and exporting data).  I'll add a note to the team about updating the documentation.

I've a couple of images (from a few years ago however I ran a quick API test to verify).  Test case is a bouncing ball and a HertzMindlin contact model.  This is the API version of HM so the unsymmetrical forces can be included/excluded.  EDEM default is 'damping not reported'.

 

 

 

 

Stephen Cole

Yu Mu said:

Hello, I still have a question about this: when sliding occurs, the tangential force calculated by edem is sliding friction. Will this force be reported? Or is it not reported as damping force?

Hi Yu,

For sliding/tangential forces this follows the same convention.  There is a damping component of the tangential force calculated during sliding which is used to determine the motion but not included when reported/exported from the Analyst.

Regards

Stephen

Yu Mu

Stephen Cole_21117 said:

Hi Yu,

For sliding/tangential forces this follows the same convention.  There is a damping component of the tangential force calculated during sliding which is used to determine the motion but not included when reported/exported from the Analyst.

Regards

Stephen

Hello Stephen,

Thank you for your patient answer, but I haven't figured out this problem yet. When browsing the H-M model code, I saw that if sliding occurs, then newF_ t will be calculated as sliding friction, and at this time F_ td will also work with newF_ t equals. Does this mean that the spring and damping components of the tangential force are replaced by a sliding friction? What I want to know is, when I derive the tangential force, will the sliding friction be derived as the tangential spring force?

Regards

Yu