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.
Help need: multi-axial object modeling
Posted 2010年7月24日 GMT+8 16:04 3 Replies
Please login with a confirmed email address before reporting spam
Hello,
I wanna modelling just like multi-transducer which are composed by two or more transducers like the tutorial model "Piezoelectrical transducer" in Acoustics model. I think that must need two or more symetric axis because each 2D symetric transducer need one. So how could make the setting to add axis in one transducer, did I need local coordinate?
Hope there are people who could supply help, I will appreciate.
Thank you.
Daniel
I wanna modelling just like multi-transducer which are composed by two or more transducers like the tutorial model "Piezoelectrical transducer" in Acoustics model. I think that must need two or more symetric axis because each 2D symetric transducer need one. So how could make the setting to add axis in one transducer, did I need local coordinate?
Hope there are people who could supply help, I will appreciate.
Thank you.
Daniel
3 Replies Last Post 2010年7月25日 GMT+8 16:20
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
Hi
one way could be to use 3D and set up a series of cylindrical coordinates (rather easy in V4) and define your physics therearound, but its true 3D so the mesh volume remains high.
Or you can make different 2Daxi geometries and couple the physics via equations (if you can easily calculate/express the coupling terms)
Probably also a few other ways, depends on how much time you have to study your case
--
Good luck
Ivar
one way could be to use 3D and set up a series of cylindrical coordinates (rather easy in V4) and define your physics therearound, but its true 3D so the mesh volume remains high.
Or you can make different 2Daxi geometries and couple the physics via equations (if you can easily calculate/express the coupling terms)
Probably also a few other ways, depends on how much time you have to study your case
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
Thank your for your reply. I tried a 3D program, but failed, because of too large dof. I am gonna simulate an array of transducers, so so large mesh I could not stand. I tried a 2D(not axial symetric), it gave me a result. But I remember that 2D geometry had a thickness of 1m, which means the shape I simulated is a rectangle rather a cylinder chip transducer.
I am thinking of using a local coordinate to the cylinder, as a user defined one. Did you think that works? If it doesn't, what situation did this one used for.
I don't know what the cylinderical coordinate system is in Coordinate System Settings of Comsol 3.5a, even afer I read the user manual. What is x coordinate of origin?
As illustrated: "Cylindrical coordinate system: A local cylindrical coordinate system (xl, yl) with origin at (x0, y0) is specified.", And I attached the diagram. If I definited (x0,y0), where is the symetric axis?
Thanks again!
Daniel
I am thinking of using a local coordinate to the cylinder, as a user defined one. Did you think that works? If it doesn't, what situation did this one used for.
I don't know what the cylinderical coordinate system is in Coordinate System Settings of Comsol 3.5a, even afer I read the user manual. What is x coordinate of origin?
As illustrated: "Cylindrical coordinate system: A local cylindrical coordinate system (xl, yl) with origin at (x0, y0) is specified.", And I attached the diagram. If I definited (x0,y0), where is the symetric axis?
Thanks again!
Daniel
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
Hi
when you use 2D and cylindrical coordinates (lets say at an origine of (x0,y0) then your axis is (x0,y0,0)-(x0,y0,1) along the Z out of plane direction. If you need several itms ou define several cylindrical coordinates, one per item, bu then you need to express also your equations this way. COMSOL looks normally after the coordinate transform.
For cylindrical axis, the default is X == phi=0 direction, if you do not give a "phi" offset too.
your coordinates are then r, phi (and a==z in 3D)
One of the difficulties in 3.5 is that you cannot "see" the coordinate system, that is you must define some vector along the different new coordinatesystem (such as a boundary force and visualise the initial conditions on the graphs to check that the coordinate system is OK, then you must take these dummy forces away as mostly they are not part of the model.
And do not forget that 2D is in fact 3D but with a depth (z) direction mostly set to 1[m] by dfault (hence simple unit transforms) but you can always change this depth value at will
--
Good luck
Ivar
when you use 2D and cylindrical coordinates (lets say at an origine of (x0,y0) then your axis is (x0,y0,0)-(x0,y0,1) along the Z out of plane direction. If you need several itms ou define several cylindrical coordinates, one per item, bu then you need to express also your equations this way. COMSOL looks normally after the coordinate transform.
For cylindrical axis, the default is X == phi=0 direction, if you do not give a "phi" offset too.
your coordinates are then r, phi (and a==z in 3D)
One of the difficulties in 3.5 is that you cannot "see" the coordinate system, that is you must define some vector along the different new coordinatesystem (such as a boundary force and visualise the initial conditions on the graphs to check that the coordinate system is OK, then you must take these dummy forces away as mostly they are not part of the model.
And do not forget that 2D is in fact 3D but with a depth (z) direction mostly set to 1[m] by dfault (hence simple unit transforms) but you can always change this depth value at will
--
Good luck
Ivar