Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Source/Destination frames used in identity pairs

Please login with a confirmed email address before reporting spam

Hi all,

I'm setting up a very simple mode of a cylinder within a rectangular domain (in assembly mode). Continuity has to be assumed on all interior boundaries and therefore, I defined an identity pair. If I now solve the model, there are some convergence problems if the Source and Destination frames of the pair are chosed to be the spatial frame. If the Source and Destination frames are chosen to be the mesh frame, the model solves correctly.

Is there anyone who can explain the difference between these frames in an identity pair. Why is one frame working and the other not?

Steven

1 Reply Last Post 3 ott 2011, 09:07 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 ott 2011, 09:07 GMT-4
Hi

you have forgotten to tell us which physics you are using, as that aslo might differentiate the frames

normally all overlap, but in "CFD" with "ale" they will all be distinguished and mostly will differ. In "solid" the spatial frame will differ from the material = mesh. For solid see it as the spatial is the deformed shape (initial + deformation u,v,w), while the reference frame is the initial shape.

Test it out: take a cube (default 1m^3), in 3D solid, stationary, give it a material property not too stiff rho = 1[kg/m^3], nu = 1/3 E = 10[Pa], add a gravity volumetric/body force (Total Force) of solid.rho*1[m/s^2], fixe one side/boundary, solve.
=> then define 2 Model Coupling Integration on the domain one on the Spatial frame, one on the Reference frame and solve the volume integration of "1" to get the volume of each.
The results, the reference volume is the initial volume, the spatial is the deformed, mostly greater

This way of decoupling by default the Spatial (x,y,z) from the Reference and Material (X,Y,Z) frame is new in v4 as is worth some careful thinking. It is basically to get CFD and Structural/Solid to look similar as its in fact two ways of looking at the same problem, with two different habit's coming from History, the story of Lagrangian and Eulerian reference frames, the one flowing with the deformation/fluid and the one onbserving from the shore ... (if that helps, take a look at en.wikipedia.org/wiki/Lagrangian_and_Eulerian_specification_of_the_flow_field ).
In this way COMSOL is pushing us to unify physics, which is of great help (in my view) when we mix all the physics.

The only exception I know about in COMSOL is the PZT material Elasticity tensor where COMSOL uses the pzt IEEE convention that differs from the usual "solid/structural" convention for the Elasticity matrix (in index ordering, be aware !!)

With the Deformed Geometry (but not the ALE) approach you need to dissociate also the Mesh frame (Xm,Ym,Zm) adding another level

So by now you should have some better clues what too choose, it all depends what you want to do ...

And why not use COMSOL to create the default pairs for you, check the corresponding tick boxes on the Geometry Assembly tab page, and check which Frame COMSOL uses ;)

--
Good luck
Ivar
Hi you have forgotten to tell us which physics you are using, as that aslo might differentiate the frames normally all overlap, but in "CFD" with "ale" they will all be distinguished and mostly will differ. In "solid" the spatial frame will differ from the material = mesh. For solid see it as the spatial is the deformed shape (initial + deformation u,v,w), while the reference frame is the initial shape. Test it out: take a cube (default 1m^3), in 3D solid, stationary, give it a material property not too stiff rho = 1[kg/m^3], nu = 1/3 E = 10[Pa], add a gravity volumetric/body force (Total Force) of solid.rho*1[m/s^2], fixe one side/boundary, solve. => then define 2 Model Coupling Integration on the domain one on the Spatial frame, one on the Reference frame and solve the volume integration of "1" to get the volume of each. The results, the reference volume is the initial volume, the spatial is the deformed, mostly greater This way of decoupling by default the Spatial (x,y,z) from the Reference and Material (X,Y,Z) frame is new in v4 as is worth some careful thinking. It is basically to get CFD and Structural/Solid to look similar as its in fact two ways of looking at the same problem, with two different habit's coming from History, the story of Lagrangian and Eulerian reference frames, the one flowing with the deformation/fluid and the one onbserving from the shore ... (if that helps, take a look at http://en.wikipedia.org/wiki/Lagrangian_and_Eulerian_specification_of_the_flow_field ). In this way COMSOL is pushing us to unify physics, which is of great help (in my view) when we mix all the physics. The only exception I know about in COMSOL is the PZT material Elasticity tensor where COMSOL uses the pzt IEEE convention that differs from the usual "solid/structural" convention for the Elasticity matrix (in index ordering, be aware !!) With the Deformed Geometry (but not the ALE) approach you need to dissociate also the Mesh frame (Xm,Ym,Zm) adding another level So by now you should have some better clues what too choose, it all depends what you want to do ... And why not use COMSOL to create the default pairs for you, check the corresponding tick boxes on the Geometry Assembly tab page, and check which Frame COMSOL uses ;) -- Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.