Discussion Forum

Hello,

I'm trying to simulate a double particle trace of ions and electrons in a volume to basically simulate ambipolar diffusion, like in plasmas. In a cylindrical volume, in one end I have the ions starting out and electrons in the other end. I would like to have these two sets of particles attract each other.

From the equations used for particle-particle interaction (which work perfectly if all the particles are doing is interacting with the same species), we have this equation built into COMSOL.

nojac(cpt.cptop1(-e_const^2*cpt.ke*(qx-dest(qx))/(1.0E-40+cpt.r^3)))

with cpt.r = sqrt((qx-dest(qx))^2+(qy-dest(qy))^2+(qz-dest(qz))^2+if(t==0,eps,1.0E-50))

And similarly there were equations for y and z.

My thought was if I created another dependent variable, where the dest(qx) term was replaced with dest(mod1.cpt2.q2x) etc. to track the locations of all the particles of the SECOND particle trace in relation to the first particle trace:

cpt.rOps = sqrt((qx-dest(mod1.cpt2.q2x))^2+(qy-dest(mod1.cpt2.q2y))^2+(qz-dest(mod1.cpt2.q2z))^2+if(t==0,eps,1.0E-50))

and likewise another Particle-Particle interaction with a user defined force:

nojac(cpt.cptop1(e_const^2*cpt.ke*(qx-dest(mod1.cpt2.q2x))/(1.0E-40+cpt.r^3))) (for all 3 dimensions)

and do the same thing for the second particle trace, but coupling it back to the first one instead, I could simulate an attractive force between the two separate particle traces.

I tried this and of many of the error messages I've received, I have this as my latest:
Failed to evaluate variable Jacobian.
- Variable: mod1.q2y
- Geometry: 1
- Domain: 2 3
Failed to evaluate operator Jacobian.
- Operator: env_pgeom_cpt
- Geometry: 2
- Boundary: 1
Failed to evaluate operator Jacobian.
- Operator: env_pgeom_cpt
- Geometry: 2
- Boundary: 1

Any ideas, tips, directions, solutions are appreciated or is it possible to cross-reference/couple particle traces in COMSOL? Much appreciated, model is attached.