| Author | Message | ||||
     mjkennedy New member Username: mjkennedy Post Number: 1 Registered: 08-2010 |
I am solving for the 2-D motion of a particle suspended in a fluid medium and under the influence of an electric field. I have time-dependent equations of motion for the particle: dt(particle_velocity) = func(E_field, fluid_velocity) dt(particle_position) = particle_velocity E_field and velocity_fluid do not depend on time. What is the best way to couple my steady state solutions for E_field and fluid_velocity into the time-dependent problem for particle_velocity? | ||||
| rgnelson Moderator Username: rgnelson Post Number: 1392 Registered: 06-2003 |
Use TRANSFER to export the field and velocity from the steady state problem, and import them into the particle system using TRANSFERMESH. | ||||
| mjkennedy New member Username: mjkennedy Post Number: 2 Registered: 08-2010 |
I wrote a simple program to compute fluid_velocity, a vector field, over a mesh for a rectangular channel. Then, I wrote a program to import this mesh and then compute particle_position, a global variable, at subsequent times for some initial position. However, I get a Singular Matrix Error on attempting to run the latter program. What might cause this?
| ||||
| rgnelson Moderator Username: rgnelson Post Number: 1395 Registered: 06-2003 |
FlexPDE assumes that it is supposed to calculate a set of scalar fields over the defined domain. These fields could be temperature or components of a velocity, or any number of things. But the position of a particle moving through the domain is not in any way I can see representable by scalar fields. You could compute flow trajectories, meaning flow from every possible incoming point. But how this is to be related to delay time from entry is beyond me. By trying to use an EVAL to limit the field to a single input point, you have merely created an under-specified system. And FlexPDE has no built-in particle-tracking facilities. |
Compute fluid_velocity