Discussion Forum

Hi Taylor, this is quite easy to do. You need to have 2 studies in the model. The first Study should be Frequency Domain, where you solve for only the fields. The second study should be a Transient study, where you only solve for the particles. A couple of notes:

1. In the Electric Force feature, you will need to use the "Multiply force by phase angle" checkbox. This tells COMSOL to compute the force value at the correct phase of the RF cycle when computing the particle trajectories.
2. You need to feed in the solution from the first study into the second. This is done in the Study settings.

For examples of how to do this, see any of these models:
www.comsol.com/model/particle-tracing-in-a-quadrupole-mass-spectrometer-10653
www.comsol.com/model/quadrupole-mass-filter-12039
www.comsol.com/model/ion-funnel-15391

OK... I am posting this still because I think it can be educational for someone else who may run across it. My problem below was not setting enough time discretization in the time dependent study. I mistakenly believed that the time step was just an "update" or "reading" and you still got the particle path between updates. That is apparently not so (learning experience for me), as soon as I made the time discretization much smaller, BAM, wobbles. I was essentially aliasing my run by only sampling at the same phase angle each time... thus it looked like a straight line.

----------------------------------------------------------------------------------------------------------------------------------------

Thank you, that is exactly what I was looking for. It is a great example for what I wish to do.

First: I would like to check on the difference between currents and electrostatics when solving for the terminals. I am not inducing a magnetic field, just a time evolving (but instantaneously solvable) static field. However, I am starting to get the itch in the back of my head that says that electrostatics cannot really be solved in the frequency domain, even though it is an option. For whatever reason the university bought electrostatics and particle tracking but not currents, so for the last day I fiddled and tried to make due with just electrostatics. Since I am duplicating everything else the same, but not able to get an oscillating force, I figure that must be the issue. I am attaching my model file for reference.

Second: I have created two parallel plates in 2D. At the moment I have set a constant voltage and an oscillating voltage in two different electrostatic physics (see above for why). I then solve the stationary ES as step one and feed that solution into the second ES as step two, a frequency domain solution. I checked the particle tracking and have applied the forces (separate) to the entire domain of interest and made sure to apply the phase angle as mentioned. I get solutions for both electric fields and as far as I can understand from the description of the physics packages it should work. When I have a non-zero force on the static field, it moves just fine. I get a nice parabola until it whacks into the plate. Next I zero the static force and apply the oscillating force to the static force (when I have it on) but doesn’t respond at all to the oscillating force. I noticed the change in the "Equation" to state that its F=eZexp(i omega t). It all seems to make sense, but no wobble.

Third: The plates are 100 m long and I release the particle at 100 m/s, for a total of 1 s of travel time. Then I set the wobble frequency to 10 Hz to get 10 wobbles in the flight path. The voltage is 0.01 V (I think) which is enough to slam it into the plate in only a fraction of the distance, which means I should see the perturbation in the flight path. Now I am stuck.

T.C. Lilly

P.S. Not stuck, the attached works just fine with two electrostatics instead of the currents model used in the example.

Hello,

I would like to do something similar, but with fluid flows.

I also use 2 studies in my model.

First, I calculate the flow, which is periodic: I simulate blood flows in diseased arteries, so there is the cardiac cycle.

Second, I use particle tracing to calculate the residence time of blood particles near the diseased sites.

Is it possible to tell Comsol that the flow (that I calculated numerically) repeats itself after one cycle for the particle tracing?

Thank you very much in advance,

Adriane