How to model magnetic pull? 2

[friedmind]

Hi All,
I'm trying to model a magnetic interaction between a shaft an the stator. The magnetic interaction behaves like a spring with negative stiffness: When the shaft approaches the stator, the attraction force increments.
I've tried using a connector element with negative stiffness, but Abaqus gives a error "ELASTIC STIFFNESS MUST BE POSITIVE".

The step is "Steady-state dynamics, Direct"

anyone knows how can I model this interaction?
Thank you in advance

 

[friedmind]

no one?

 

[rstupplebeen]

I have never tried however here is a guess. You could vary a pressure load based on the relative distance between nodes. If you provide more details into what you are trying to model, geometry and desired outcomes we will probably be able to help more.

Rob Stupplebeen

https://sites.google.com/site/robertkstupplebeen/

 

[CharlieM80]

Do you mean like a magnetic bearing? Those are unstable systems, and need an 'active' external PID control system to enforce stability. AFAIK, you cannot use FEM with such unstable systems alone.

If you are trying to characterize a system with a magnetic bearing, the bearing will have a positive stiffness when you look at the whole system with the active controller.

If you mean something else, such as a shaft supported normally with bearings, but subject to a magnetic field body force, then I guess that is possible. Not sure how to do that though.

Are you looking to include the damping effect from induced eddy currents?

 

[friedmind]

Thanks rstupplebeen and CharlieM80 for your answers. What I'm trying to model is a Kaplan turbine shaft. It's a vertical shaft supported by two bearings and subjected to a magnetic body force on the top end (generator). This body force is attractive, depending on the distance between rotor and stator. I'm not looking to include the damping effect from induced eddy currents at this point.

I guess I could try the approach of rstupplebeen, but I don't know how to implement it.
Do I have to use DLOAD to vary a pressure load (or body force) based on the relative distance between nodes?
Can this be used on a "Steady-state dynamics, Direct" step?

Thank you very much for your help.

matias