Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

problem with EDEM on Linux

Hey

I am trying to make a plate fall from a height of 0.0105m with an initial velocity of 20 m/s. The time required for the plate to reach the ground is 0.5 milliseconds.

If i run the simulation in Windows and give this in the environment with this as motion defination

Gravity will be applied on the plate in -ve z.

The gravity vector will be applied to the plate as it falls down.

Now, if I do the same thing on Linux and do not give acceleration in motion definition,

It will not work. It was working only when I defined a gravity acceleration in the acceleration part of the motion

Why so?

Is this a bug in EDEM, or am I doing anything wrong here? Please let me know.

In both Windows and Linux i am using EDEM2025.

Find more posts tagged with

EDEM

v2025.0

Accepted answers

All comments

Stephen Cole

HI, the user interface on Windows and Linux is the same.

I'm not sure from the description what the issue is? If you setup the model in Windows and run on Linux does it run OK?

Regards

Stephen

satri

Running is not a problem. The issue is gravity is not applied to the geometry even after specifying it in environment for Linux. Only particle experience gravity. If I give motion to geometry it will not experience gravity unless acceleration is specified in motion. I was under the impression that if I specify acceleration for motion as well as gravity then the geometry Will experience double acceleration.

This is not the case in windows. Even if I don't specify the acceleration the geometry Will experience gravity as it moves downward. Because it's activated in environment. But this doesn't happen on Linux.

Stephen Cole

The gravity set in the Creator> Environment applies to particles only, same for Windows and Linux.

If you specify a kinematic you can 'include gravity in force calculation with a Motion Controller, otherwise gravity won't influence geometry motion.

satri

That option you have highlighted does not show up in Linux.

I am using edem2025. Is there anything that needs to be done to activate that option?

satri

The image of motion specification i have put above is from Linux. I do not see that option there

Stephen Cole

Hi,

Only the Force Controller has the option to include gravity. I think the image you shared is from a Linear Translation and there isn't the option to add gravity force on Windows or Linux .

Regards

Stephen

satri

But please try this: drop a geometry from a particular height. You can estimate the time by solving the quadratic eq for distance s=ut+0.5gt^2 where u is initial velocity, t is time, s is displacement and g is gravity.

If you activate gravity and not specify acceleration in linear translation motion in windows machine but activate gravity in environment then the geometry Will reach the ground at the solution time of quadratic equation.

But if you solve the exact same model in Linux the geometry Will not reach the ground in the time specified by quadratic eq.

Now in the Linux machine if you specify the acceleration in linear translation motion then the geometry Will reach the ground at the time specified by solution of quadratic equation.

Please let me know if you do not observe the same thing.

Stephen Cole

Without specifying acceleration with Linear Translation we see the linear displacement only, no influence of gravity on Windows or Linux. Are you running at the same time-step/save intervals on Windows and Linux?

Regards
Stephen

satri

Maybe i should rephrase my statement. Its not problem with Windows or Linux. I notice this strange behaviour in windows as well as linux please tell me why this is happening

my time step was 2.5e-10s and initial velocity of the plate was 0.1m/s in -ve z direction save interval was always 0.0001s

The plate started at a height of 10100 micrometers (0.0101m). and ran for time duration of 0.035s when i give it no acceleration in motion definition the plate stopped at a height of 6600 micrometers (6.6e-3m) from the ground.

solution to quaderatic eq

(0.5*g)t^2+0.1t-(10100-500)*10^(-6) = 0 solution to t= 0.035,-0.055, neglect -0.055. time of flight is 0.035s

I did not run the simulation with the gravity value given in the motion definition for this time step. because it takes two days to run.

insted i did the following

simulation A (Linux)

I give time step of 2.5e-6s and initial velocity of 0.1m/s in -ve z direction and no acceleration is definied in the motion defination (gravity is activated in the environment). The plate stops at the height of 6600 micrometers (6.6e-3m) above the ground after a time of 0.035s

simulation B (Linux)

I give a time step of 2.5e-6s and an initial velocity of 0.1m/s in the -ve z direction, and acceleration=9.81 m/s^2 is defined in the motion definition (gravity is activated in the environment). The plate stops at the height of 500 micrometer (0.5e-3m) above ground after a time of 0.035s.

simulation C (Windows)

I give a time step of 2.5e-6s and an initial velocity of 0.1m/s in the -ve z direction, and no acceleration is defined in the motion definition (gravity is activated in the environment). The plate stops at the height of 500 micrometer (0.5e-3m) above ground after a time of 0.035s.

so there is no problem with windows and linux. but why is gravity not acting on the geometry?

Because when i give the initial velocity of 20m/s the time of flight is 0.00048 and the plate does not stop at such a large height.

so, what is the correct way to go about? if i want the plate to come and stop at a distance of 500 micrometers above the ground?

Stephen Cole

Hi,

Just for clarification gravity as defined in the Creator > Environment setting only acts on geometry if a Force Controller is setup to use this. For Kinematics the gravity never influences the motion.

Simulation A and B act as I would expect, although it should end up 591.4 microns from the ground rather than 500 in B. You'd need 0.035206 s run time to be closer, and to be exactly 500 microns a smaller time-step.

Simulation C isn't behaving as expected, if this is a linear translation kinematic with no acceleration it will just move according to the linear velocity set as you have in Simulation A. I would expect there is something else happening in C, such as additional kinematics set or running at a different time-step, inclusion of parent-child kinematics?

satri

I see. your first sentence explains why its stopping in between

you are right about simulation A and B. i only put the rounded numbers.

i dont want to use force controller here. i am using linear translation. Your first statement clarifies what I want to know. i will define acceleration if i want gravity to act.

please tell me this,

If i activate gravity in the environment gravity will be activated by default as body force for particles right. (just checking)

Stephen Cole

Hi, yes if you select gravity in the environment section it gets applied to all particles as default, no additional models required.