Discussion Forum

Hi,

You can leave the sub-domain on "free displacement".
Normally it should work since that you have the acceleration. You can set a velocity to the concerned boundaries, which should be connected to the acceleration via the time stepping.

Good luck ,

Hi again,

Thanks a lot for your replay! It sounds good... do you possibly have any more information about what variables to use?

I guess that the formulas for the boundary conditions shall look something like:
v = v0 + a * deltaT
where v0 is the velocity in the previous time step, a is the acceleration and deltaT is the time step length.

or if the displacement is specified:

x = x0 + v0 * deltaT + (1/2) * a * deltaT^2 (linear motion)

where x0 is the displacement (=position) in the previous time step. But what are the variable names in COMSOL needed to obtain the values of x0, v0 and deltaT?

Regards,
Johan Gustafsson

Hi,

It is much better to set velocity than displacement, but it should work for both.
I have attached a simple model (that I have already posted for practically the same purpose) which can help you to understand more.
The model is composed of two parallel plates. One in the bottom fixed and the second in the top supposed to move with a certain velocity.
I have set the velocity dependent with time linearly (in global expressions):

V=acceleration*t With an adequate time range and step.

Note that the value representing the acceleration in this model is negative. It does not mean a deceleration but it means the direction of the displacement negative or positive following your displacement. This is a convention when you set velocities, you have to take into account the direction of the displacement.

I hope it helps and good luck,

Cheers

Hi again,

Thanks a lot for helping me by uploading the example model and the tips of using the boundary conditions, unfortunately I still haven’t been able to completely resolve the issue.

The plate_motion.mph example has, like you say, a constant acceleration specified. But in my model the acceleration depends on the displacement and velocity of the subdomain. The boundary conditions for the moving mesh application mode only allow the mesh displacement or the mesh velocity to be specified, not the mesh acceleration unfortunately.

I need to somehow integrate the velocity due to the acceleration. Similar to this:
v = v0 + a * deltaT

where v0 is the velocity in the previous time step, a is the acceleration and deltaT is the time step length. As I understand it the time variable (t) is just a scalar variable for the current time (not the time step?). I guess that the current velocity v0 can be obtained as “zt”, the time derivative in the z-direction, (my model is axisymmetric). But I have still have not been able to figure out how to express the integral of the acceleration.

Thanks,
Johan

interrresting problem you are investigating
why dont you use a global equation to calculate the velocity from acceleration?
if you have gamma you just solve vt-gamma=0

JF

Hi,

I could not understand what you are trying to do with this integration!!!

In your first post, you have mentioned that your acceleration is calculated within the Maxwell stress tensor!!
So I really could not understand this : "But in my model the acceleration depends on the displacement and velocity of the sub-domain"??
If your acceleration is time dependent it is not a problem. Just replace it in the previous equation and that's all!!

I hope it helps,

Cheers,

Thanks a lot R. Faycal and Jean Francois, for taking time to discuss this problem with me. It has been very helpful. Specifying a global equation did the trick. The reason that I confused you regarding my problem is that to start with I only wanted to specify the acceleration caused by an EM force, but in my problem I also have a displacement dependent force (a spring) and a velocity dependent (damping). So I ended up having to specify a global equation to solve for the displacement and not the velocity and it seems to work fine.

Thanks again,
Johan

Dear Sir,
Can you tell me how to write global equation in differential and integral form. thank you in advance
and regards

hi
\i am also trying same kind of model. In my case, forces acting on the armature (moving part)are magnetic and gravitational forces
> when current of the electromagnet is switched off, the armature falls due to its weight.
> when i am using ALE, the armature is moving but magnetic flux density in armature is remaining same, on the contrary, the value of B should decrease.
> in starting i am not switching of my current. and armature is moved with a fixed velocity.

Please give some suggestion.
thanks
vijay sharma

Hi Johan,

I have the same problem as you had. I am studying the movement of the armature of an electromagnetic actuator and I used the magnetic fields physics to calculate the magnetic force. After that I defined the total force at the global equations physics (the forces acting in the plunger are the magnetic force plus the spring force). The problem I have at the moment is that the armature moves more than it should and interferes with the fixed parts, so I was assuming that I do something wrong in the definition of the forces.

Can you share a screenshot of how did you setup the Newton's Law of Motion and the displacement if possible?

Kind regards,
Pavlos.