Reactionary forces between geometry and particles
Hello,
I've just begun using EDEM, and I'm having trouble understanding when reaction forces are captured. I have a model where an imported geometry collides with particles. I'd like to include some initial movement conditions on the geometry. I see in the documentation it says,
Unlike kinematics, you can add motion controllers to the geometry elements to allow objects to react to particle-geometry forces. Forces acting on the geometry are included in the calculation of the geometries motion. There is no contact detection between geometry elements. Sections pass through each other when they come into contact.
However, when I add a torque controller with an initial angular velocity, I don't see the reactionary forces I expect. In particular, the geometry does not move at all at the location where the initial angular velocity is applied. Does the torque controller create a joint at that location such that translation is prevented?
Thanks.
Best Answer
-
bocaj22 said:
Hi Stephen,
Thank you for sharing this, it acts as I suspected. Is there a way to initialize movement without adding a constraint like this? Perhaps it doesn't make sense to be able to do this from any location, but I think I should be able to initialize some angular momentum or angular velocity such that the geometry is rotating about its center of gravity, but still able to translate freely when impacted by a particle.
Thanks,
Bocaj22
Hi,
You can have more than 1 controller although the force/torque control is designed for fairly simple motion types. For example if in the torque controller you set the rotation around the centre of mass and 'moves with body' selected you can also create a force controller which if you set 0 force the geometry would translate due to the particle force and always rotate around the centre of mass if this rotation point is moving with the body.
However this does depend on the complexity of the motion required, you could consider coupling with MotionSolve for more complex motions - https://help.altair.com/hwsolvers/ms/topics/tutorials/mv/tut_mv_7021_ms_edem_cosimulation_t.htm#task_vfc_nfm_kjb
Regards
Stephen
0
Altair EmployeeAnswers
-
Hi,
If you have a force or torque controller then the force from the particle on the geometry can influence it's motion. With the force/torque controller you set the mass of the geometry and also can constrain the direction in which it rotates or moves. It maybe that either the geometry mass is such that the particle force has negligible effect on the motion or that any constraints set for the motion means the direction of the applied force doesn't cause a rotation.
You can see more on the setup of Motion Controllers in the e-learning lesson 5.4 https://learn.altair.com/course/view.php?id=147
Regards
Stephen
0 -
Stephen Cole_21117 said:
Hi,
If you have a force or torque controller then the force from the particle on the geometry can influence it's motion. With the force/torque controller you set the mass of the geometry and also can constrain the direction in which it rotates or moves. It maybe that either the geometry mass is such that the particle force has negligible effect on the motion or that any constraints set for the motion means the direction of the applied force doesn't cause a rotation.
You can see more on the setup of Motion Controllers in the e-learning lesson 5.4 https://learn.altair.com/course/view.php?id=147
Regards
Stephen
Hi Stephen,
Thank you for your reply and link. Unfortunately the video did not completely answer my question. Without any constraints applied elsewhere, and without the "Constrain to 1D Rotation" checked, should the initial angular velocity setting of the torque motion controller be creating a joint constraint at the point of action (revolute or spherical)? Perhaps I need to switch to a torque equation instead?
Thanks,
bocaj22
0 -
bocaj22 said:
Hi Stephen,
Thank you for your reply and link. Unfortunately the video did not completely answer my question. Without any constraints applied elsewhere, and without the "Constrain to 1D Rotation" checked, should the initial angular velocity setting of the torque motion controller be creating a joint constraint at the point of action (revolute or spherical)? Perhaps I need to switch to a torque equation instead?
Thanks,
bocaj22
Hi,
Just creating the torque controller creates the constraint around the point of action. Without the "constrain to 1D rotation" it will rotate around this point in any direction. The initial angular velocity just sets it rotating around this point at the specified velocity but doesn't influence how/where the torque is applied.
I created a quick example with a box initially rotating around the Y axis and Manual Point of Action set to be at the base of the box not at the centre of mass. Initially it rotates around this point in the Y axis until it is hit by the particles at which point it rotates in all 3 axis due to the particle force (as no force is set to apply the initial motion, just initial velocity) but still around this point.
Images below show the setup. The image on the right is just showing the points in the Analyst just as reference, not the rotation axis.
0 -
Stephen Cole_21117 said:
Hi,
Just creating the torque controller creates the constraint around the point of action. Without the "constrain to 1D rotation" it will rotate around this point in any direction. The initial angular velocity just sets it rotating around this point at the specified velocity but doesn't influence how/where the torque is applied.
I created a quick example with a box initially rotating around the Y axis and Manual Point of Action set to be at the base of the box not at the centre of mass. Initially it rotates around this point in the Y axis until it is hit by the particles at which point it rotates in all 3 axis due to the particle force (as no force is set to apply the initial motion, just initial velocity) but still around this point.
Images below show the setup. The image on the right is just showing the points in the Analyst just as reference, not the rotation axis.
Hi Stephen,
Thank you for sharing this, it acts as I suspected. Is there a way to initialize movement without adding a constraint like this? Perhaps it doesn't make sense to be able to do this from any location, but I think I should be able to initialize some angular momentum or angular velocity such that the geometry is rotating about its center of gravity, but still able to translate freely when impacted by a particle.
Thanks,
Bocaj22
0 -
bocaj22 said:
Hi Stephen,
Thank you for sharing this, it acts as I suspected. Is there a way to initialize movement without adding a constraint like this? Perhaps it doesn't make sense to be able to do this from any location, but I think I should be able to initialize some angular momentum or angular velocity such that the geometry is rotating about its center of gravity, but still able to translate freely when impacted by a particle.
Thanks,
Bocaj22
Hi,
You can have more than 1 controller although the force/torque control is designed for fairly simple motion types. For example if in the torque controller you set the rotation around the centre of mass and 'moves with body' selected you can also create a force controller which if you set 0 force the geometry would translate due to the particle force and always rotate around the centre of mass if this rotation point is moving with the body.
However this does depend on the complexity of the motion required, you could consider coupling with MotionSolve for more complex motions - https://help.altair.com/hwsolvers/ms/topics/tutorials/mv/tut_mv_7021_ms_edem_cosimulation_t.htm#task_vfc_nfm_kjb
Regards
Stephen
0 -
Stephen Cole_21117 said:
Hi,
You can have more than 1 controller although the force/torque control is designed for fairly simple motion types. For example if in the torque controller you set the rotation around the centre of mass and 'moves with body' selected you can also create a force controller which if you set 0 force the geometry would translate due to the particle force and always rotate around the centre of mass if this rotation point is moving with the body.
However this does depend on the complexity of the motion required, you could consider coupling with MotionSolve for more complex motions - https://help.altair.com/hwsolvers/ms/topics/tutorials/mv/tut_mv_7021_ms_edem_cosimulation_t.htm#task_vfc_nfm_kjb
Regards
Stephen
Okay, that makes sense. I appreciate the help.
0