Menu

Discussion Forum

New Discussion
Filtrer

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.

magnetic force for structure bending

Posted 4 sept. 2009, 02:22 UTC−4 27 Replies

Shu Zhou
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi I was given a task to show how a polymer bends. Basically I put a Nickel just on the polymer then a strong magnet underneath. There is space bvetween the magnet and polymer so the polymer is supposed to bend down because of the attraction force between the Nickel and magnet.
I am using the two modules Magnetostatics, No Current and Solid, Stress-Strain. I tested it with the magnetostatics modules only about the magnetic flux. The magnextic flux did go through the Nickel. But later on I added the solid stress module to see the bending. But nothing happened, the total displacement is 0. I am wondering if you guys know the reason?
I use a big box to cover all of them and put air in the box since I read the tutorial models " Permanent Magnet". I think it should be a simple model since once you have the force on Nickel it should at least changes in shape....

Thanks for helping!
27 Replies Last Post 26 sept. 2009, 17:58 UTC−4
Shu Zhou
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 4 sept. 2009, 17:07 UTC−4
Anyone knows? Please help! Thanssssss....
Anyone knows? Please help! Thanssssss....
Jigar Patel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 7 sept. 2009, 19:19 UTC−4
I am having the same issue. I am putting an iron particle in a soft material, and then applying a magnetic field to it. The flux lines and the field all seem to be working properly; however, the particle does not move at all. It doesn't make sense. Did you try using both the stress and strain modules? Or were you able to figure it out? If so, I would appreciate any help you could provide.

Thanks
I am having the same issue. I am putting an iron particle in a soft material, and then applying a magnetic field to it. The flux lines and the field all seem to be working properly; however, the particle does not move at all. It doesn't make sense. Did you try using both the stress and strain modules? Or were you able to figure it out? If so, I would appreciate any help you could provide. Thanks
Shu Zhou
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 9 sept. 2009, 00:23 UTC−4
Nope, I am still waiting. I was using solid, stress-strain module. But it doesn't work
Nope, I am still waiting. I was using solid, stress-strain module. But it doesn't work
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 9 sept. 2009, 03:06 UTC−4
I think the problem is that you have not actually coupled the solid stress-strain application mode with the magnetostatics application mode. To couple them, you have to put the lorentz forces on the struture. To do that , when the solid stress-strain application mode is active, go to the subdomain tab and in the load tab enter FLtzx_emqa, FLtzy_emqa and FLtzz_emqa for Fx, Fy and Fz ,respectively. If you don't do this, COMSOL does not understand that you want to couple the magnetstatic mode with solid stress-strain mode. If you want to know the details about FLtzx_emqa, FLtzy_emqa and FLtzz_emqa , go to physics?equation system/subdomain settings/variables and find the Lorentz force in the description.
I hope it helps
Nastaran
I think the problem is that you have not actually coupled the solid stress-strain application mode with the magnetostatics application mode. To couple them, you have to put the lorentz forces on the struture. To do that , when the solid stress-strain application mode is active, go to the subdomain tab and in the load tab enter FLtzx_emqa, FLtzy_emqa and FLtzz_emqa for Fx, Fy and Fz ,respectively. If you don't do this, COMSOL does not understand that you want to couple the magnetstatic mode with solid stress-strain mode. If you want to know the details about FLtzx_emqa, FLtzy_emqa and FLtzz_emqa , go to physics?equation system/subdomain settings/variables and find the Lorentz force in the description. I hope it helps Nastaran
Shu Zhou
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 9 sept. 2009, 04:27 UTC−4
Hi Nastaran
Thank you so much for your help. In my case actually I used Magnitostatics, No current (emnc) module. So do I need to change it to Magnitostatics, Vector potential (emqa) instead as your suggestion? Since I am modeling the force between two magnet (without current) so I get confused about it. for emnc, I have tried to put the electromagnetic force (in my case is Nickel_forcez_emnc) on the structure. Then I can't get any solution (maybe it is not convergent)...so any idea on that?

Thank you again for your help!

Shu
Hi Nastaran Thank you so much for your help. In my case actually I used Magnitostatics, No current (emnc) module. So do I need to change it to Magnitostatics, Vector potential (emqa) instead as your suggestion? Since I am modeling the force between two magnet (without current) so I get confused about it. for emnc, I have tried to put the electromagnetic force (in my case is Nickel_forcez_emnc) on the structure. Then I can't get any solution (maybe it is not convergent)...so any idea on that? Thank you again for your help! Shu
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 9 sept. 2009, 04:47 UTC−4
Actually I think you need to change to Magnitostatics, Vector potential (emqa) instead. That is what I did and it worked for my case(permanent magnet force on a metal beam without current) . Actually in case of Magnitostatics, No current (emnc) module when I look at physics/equation systems/subdomain settings/variables, I don't see any variable for loretnz force.The same is in the postprocessore section.Maybe I am missing it. I don't know the reason. But I see Lorentz force in case of Magnitostatics, Vector potential (emqa) . So it means that the Lorentz force is not defined in the Magnitostatics, No current (emnc) module. right?
Actually I think you need to change to Magnitostatics, Vector potential (emqa) instead. That is what I did and it worked for my case(permanent magnet force on a metal beam without current) . Actually in case of Magnitostatics, No current (emnc) module when I look at physics/equation systems/subdomain settings/variables, I don't see any variable for loretnz force.The same is in the postprocessore section.Maybe I am missing it. I don't know the reason. But I see Lorentz force in case of Magnitostatics, Vector potential (emqa) . So it means that the Lorentz force is not defined in the Magnitostatics, No current (emnc) module. right?
Shu Zhou
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 9 sept. 2009, 05:34 UTC−4
Also, some idea just flash in my head and I guess I can try to put the maget force as a global expression on structure. I gonna try it out.


Also, some idea just flash in my head and I guess I can try to put the maget force as a global expression on structure. I gonna try it out.
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 9 sept. 2009, 05:42 UTC−4
Great!
Just let me know if it works with emqa or emnc ?
Nastaran
Great! Just let me know if it works with emqa or emnc ? Nastaran
Shu Zhou
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 9 sept. 2009, 07:19 UTC−4
Nastaran,
Bad news! It is not working with both emqa and emnc. I misput some non-zero value in Jx, which is supposed to be zero in my case. I guess in my case the loretnz force is zero. But in your model, you have moving metal which is like a moving charge. So you have current then you have lorentz force....

I think my key problem is in comsol I don't know how to calculate the magnetc force for a magnet in a static magnetic field . Otherwise I can put this force instead of FLtzz_emqa into Fz....
Nastaran, Bad news! It is not working with both emqa and emnc. I misput some non-zero value in Jx, which is supposed to be zero in my case. I guess in my case the loretnz force is zero. But in your model, you have moving metal which is like a moving charge. So you have current then you have lorentz force.... I think my key problem is in comsol I don't know how to calculate the magnetc force for a magnet in a static magnetic field . Otherwise I can put this force instead of FLtzz_emqa into Fz....
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 9 sept. 2009, 21:55 UTC−4
Shu,
You are right!Lorentz force is not zero for your case. You said that you have studied the permanent magnet example in the manual. I read it too so it seems that the Magnetostatics, No Currents application mode should also work for you. Did you read this lines in the manual which says: In the manual it is written that “In magnetostatic problems, where no currents are present, the problem can be solved using a scalar magnetic potential. This model illustrates this technique for a horse shoe-shaped permanent magnet placed near an iron rod.” Also it says “To calculate the force on the rod, use the surface stress tensor, which is derived in “Electromagnetic Forces” on page 133 in the AC/DC Module User’s Guide, and “To evaluate the force in the x direction, open the Global Data Display dialog box and enter the expression rod_forcex_emnc.” I tried it and in the Data Display /Global/ I can see the magnitude of force. An approximate method could be that you apply the 3 components of rod_forcex_emnc, y and z as body forces in Solid stress-strain application mode, you apply this force as a body force you have to divide them to the volume of the object and apply them as N/m^3. But this means that you have to assume that force distribution is uniform in the volume of the object. I guess there should be a better method by coupling surface stress tensor to solid strain-stress mode but right now I don’t know how.I think COMSOL knows how to calculate the magnetc force for a magnet in a static magnetic field but it gives it as a total force "rod_forcex_emnc" so it is not easy to apply it in solid stress strain mode. I think we need to know how to use surface stress tensor.
By the way sorry for misleading you yesterday.
Nastaran
Shu, You are right!Lorentz force is not zero for your case. You said that you have studied the permanent magnet example in the manual. I read it too so it seems that the Magnetostatics, No Currents application mode should also work for you. Did you read this lines in the manual which says: In the manual it is written that “In magnetostatic problems, where no currents are present, the problem can be solved using a scalar magnetic potential. This model illustrates this technique for a horse shoe-shaped permanent magnet placed near an iron rod.” Also it says “To calculate the force on the rod, use the surface stress tensor, which is derived in “Electromagnetic Forces” on page 133 in the AC/DC Module User’s Guide, and “To evaluate the force in the x direction, open the Global Data Display dialog box and enter the expression rod_forcex_emnc.” I tried it and in the Data Display /Global/ I can see the magnitude of force. An approximate method could be that you apply the 3 components of rod_forcex_emnc, y and z as body forces in Solid stress-strain application mode, you apply this force as a body force you have to divide them to the volume of the object and apply them as N/m^3. But this means that you have to assume that force distribution is uniform in the volume of the object. I guess there should be a better method by coupling surface stress tensor to solid strain-stress mode but right now I don’t know how.I think COMSOL knows how to calculate the magnetc force for a magnet in a static magnetic field but it gives it as a total force "rod_forcex_emnc" so it is not easy to apply it in solid stress strain mode. I think we need to know how to use surface stress tensor. By the way sorry for misleading you yesterday. Nastaran
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 10 sept. 2009, 01:30 UTC−4
I correct my first sentence: Lorentz force is zero in your model.
Sorry
Nastaran
I correct my first sentence: Lorentz force is zero in your model. Sorry Nastaran
Shu Zhou
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 10 sept. 2009, 04:36 UTC−4
Hi Nastaran
Thank you for spending so much time. Actually I have tried this way today morning and it worked. As you said, the force has to be uniform......it is not perfect but give some approximate results... :)
Hi Nastaran Thank you for spending so much time. Actually I have tried this way today morning and it worked. As you said, the force has to be uniform......it is not perfect but give some approximate results... :)
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 10 sept. 2009, 06:12 UTC−4
Hi Shu
Good!
Did you also consider 3 components of Torques (rod_torquex_emnc,y,z) ?
I really think that we have to use Maxwell stress tensor for more accurate solution. I can see it in Physics/Equation system/boundary setting/variables.
There are unTx_emnc,y and z and dnTx_emnc. There are the Maxwell stress tensors on each boundary. We have to apply this stress tensor in the solid stress strain mode on the boundaries. But I don't know where to apply them.
Nastaran
Hi Shu Good! Did you also consider 3 components of Torques (rod_torquex_emnc,y,z) ? I really think that we have to use Maxwell stress tensor for more accurate solution. I can see it in Physics/Equation system/boundary setting/variables. There are unTx_emnc,y and z and dnTx_emnc. There are the Maxwell stress tensors on each boundary. We have to apply this stress tensor in the solid stress strain mode on the boundaries. But I don't know where to apply them. Nastaran
Jigar Patel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 10 sept. 2009, 19:04 UTC−4


I think the problem is that you have not actually coupled the solid stress-strain application mode with the magnetostatics application mode. To couple them, you have to put the lorentz forces on the struture. To do that , when the solid stress-strain application mode is active, go to the subdomain tab and in the load tab enter FLtzx_emqa, FLtzy_emqa and FLtzz_emqa for Fx, Fy and Fz ,respectively. If you don't do this, COMSOL does not understand that you want to couple the magnetstatic mode with solid stress-strain mode. If you want to know the details about FLtzx_emqa, FLtzy_emqa and FLtzz_emqa , go to physics?equation system/subdomain settings/variables and find the Lorentz force in the description.
I hope it helps
Nastaran


I included this in both the stress and the strain domains; however, there is an error saying that it failed to evaluate the variable. And if I leave those to zero, the simulation runs but gives me a displacement and stress of 0 which is definitely not right. Is there something I am doing wrong in defining the model? I apologize for such an open ended question but I am new to Comsol.

Thanks for your help
[QUOTE] I think the problem is that you have not actually coupled the solid stress-strain application mode with the magnetostatics application mode. To couple them, you have to put the lorentz forces on the struture. To do that , when the solid stress-strain application mode is active, go to the subdomain tab and in the load tab enter FLtzx_emqa, FLtzy_emqa and FLtzz_emqa for Fx, Fy and Fz ,respectively. If you don't do this, COMSOL does not understand that you want to couple the magnetstatic mode with solid stress-strain mode. If you want to know the details about FLtzx_emqa, FLtzy_emqa and FLtzz_emqa , go to physics?equation system/subdomain settings/variables and find the Lorentz force in the description. I hope it helps Nastaran [/QUOTE] I included this in both the stress and the strain domains; however, there is an error saying that it failed to evaluate the variable. And if I leave those to zero, the simulation runs but gives me a displacement and stress of 0 which is definitely not right. Is there something I am doing wrong in defining the model? I apologize for such an open ended question but I am new to Comsol. Thanks for your help
Shu Zhou
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 10 sept. 2009, 19:52 UTC−4
Maybe you are using the wrong module, FLtzx_emqa, FLtzy_emqa and FLtzz_emqa is for magnetostatics, Vector Potential. If you use other modules, their name is not emqa, so comsol doesn't define these variables.
Maybe you are using the wrong module, FLtzx_emqa, FLtzy_emqa and FLtzz_emqa is for magnetostatics, Vector Potential. If you use other modules, their name is not emqa, so comsol doesn't define these variables.
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 10 sept. 2009, 21:53 UTC−4
jigar,
I am also new to multiphysics in comsol and I am also here to learn. Can you explain more about your model and the application mode that you are using? I am not clear when you say "both the stress and the strain domains" There is one solid stress-strain application mode or domain and one magnetostatic application mode.As Shu said when you use a variable, make sure that it exist in that applicatiom mode by going to physics/equation system/subdomain or boundary settings/variables. Also make sure not to apply a boundary variable from one application mode to subdomain from another application mode.For example if you have current then in the physics/equation system/subdomain settings/variables you have Lorentz force values in the magnetostatic mode which is a subdomain variable with the N/m^3 unit, for applying them in the solid stress strain mode you have to consider it as a body force and not surface force.On the other hand if you go to /equation system/boundary settings/variables you can see the variables for Lorentz forces on each bounday with the unit (Pa=N/m^2) .If you want to apply them on your structure in the solid strain/stress mode you have to apply them on the boundary as surface or boundary loads.
Nastaran
jigar, I am also new to multiphysics in comsol and I am also here to learn. Can you explain more about your model and the application mode that you are using? I am not clear when you say "both the stress and the strain domains" There is one solid stress-strain application mode or domain and one magnetostatic application mode.As Shu said when you use a variable, make sure that it exist in that applicatiom mode by going to physics/equation system/subdomain or boundary settings/variables. Also make sure not to apply a boundary variable from one application mode to subdomain from another application mode.For example if you have current then in the physics/equation system/subdomain settings/variables you have Lorentz force values in the magnetostatic mode which is a subdomain variable with the N/m^3 unit, for applying them in the solid stress strain mode you have to consider it as a body force and not surface force.On the other hand if you go to /equation system/boundary settings/variables you can see the variables for Lorentz forces on each bounday with the unit (Pa=N/m^2) .If you want to apply them on your structure in the solid strain/stress mode you have to apply them on the boundary as surface or boundary loads. Nastaran
Jigar Patel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 14 sept. 2009, 13:54 UTC−4

jigar,
I am also new to multiphysics in comsol and I am also here to learn. Can you explain more about your model and the application mode that you are using? I am not clear when you say "both the stress and the strain domains" There is one solid stress-strain application mode or domain and one magnetostatic application mode.As Shu said when you use a variable, make sure that it exist in that applicatiom mode by going to physics/equation system/subdomain or boundary settings/variables. Also make sure not to apply a boundary variable from one application mode to subdomain from another application mode.For example if you have current then in the physics/equation system/subdomain settings/variables you have Lorentz force values in the magnetostatic mode which is a subdomain variable with the N/m^3 unit, for applying them in the solid stress strain mode you have to consider it as a body force and not surface force.On the other hand if you go to /equation system/boundary settings/variables you can see the variables for Lorentz forces on each bounday with the unit (Pa=N/m^2) .If you want to apply them on your structure in the solid strain/stress mode you have to apply them on the boundary as surface or boundary loads.
Nastaran


Nastaran,

I am trying to model the effects of a magnetic field on iron particles subjected in an elastic media. So using the magnetostatics module, I created/defined the permanent magnet's properties, as well as the iron particles, and the elastic media. Then I used the structural module to define the elastic properties of the media. As I see it, that's all I needed to define right? Also, how should I define the particle's geometry inside the media? I have attached a copy of the model I'm working on just in case I'm explaining the model in a wrong way. I really appreciate your help.

Thanks,
Jigar
[QUOTE] jigar, I am also new to multiphysics in comsol and I am also here to learn. Can you explain more about your model and the application mode that you are using? I am not clear when you say "both the stress and the strain domains" There is one solid stress-strain application mode or domain and one magnetostatic application mode.As Shu said when you use a variable, make sure that it exist in that applicatiom mode by going to physics/equation system/subdomain or boundary settings/variables. Also make sure not to apply a boundary variable from one application mode to subdomain from another application mode.For example if you have current then in the physics/equation system/subdomain settings/variables you have Lorentz force values in the magnetostatic mode which is a subdomain variable with the N/m^3 unit, for applying them in the solid stress strain mode you have to consider it as a body force and not surface force.On the other hand if you go to /equation system/boundary settings/variables you can see the variables for Lorentz forces on each bounday with the unit (Pa=N/m^2) .If you want to apply them on your structure in the solid strain/stress mode you have to apply them on the boundary as surface or boundary loads. Nastaran [/QUOTE] Nastaran, I am trying to model the effects of a magnetic field on iron particles subjected in an elastic media. So using the magnetostatics module, I created/defined the permanent magnet's properties, as well as the iron particles, and the elastic media. Then I used the structural module to define the elastic properties of the media. As I see it, that's all I needed to define right? Also, how should I define the particle's geometry inside the media? I have attached a copy of the model I'm working on just in case I'm explaining the model in a wrong way. I really appreciate your help. Thanks, Jigar

Attachments:

Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 14 sept. 2009, 22:33 UTC−4
Jigar
Your model is difficult for me to understand but I found following points
I-You have to use either plane stress or plane strain application mode and not both of them. In simple words, if the thickness in z direction is very small you use plane stress and model x-y plane and if the thickness in z direction is high but you want to model a cross section you consider plane strain and model x-y plane.
II-When you select a structural module like plane stress or plane stress, then you have to put supports in your model to make it structurally stable otherwise your model is not stable and FEM programs can’t deal with it.
III-You input µ =4*3.14*10^(-7) or µ =2000*4*3.14*10^(-7) in the B= µ0 µr H but 4*3.14*10^(-7) is not necessary. Because µr?is the relative permeability of the medium with respect to air’s permeability (µ0) so the minimum value for µr is 1 when the medium is air. ?0 = 4?×10?7 which is for the air exist already in the equation B= µ0 µr H.
I try to look more in your model but can you fix the above issues and send it again
Nastaran
Jigar Your model is difficult for me to understand but I found following points I-You have to use either plane stress or plane strain application mode and not both of them. In simple words, if the thickness in z direction is very small you use plane stress and model x-y plane and if the thickness in z direction is high but you want to model a cross section you consider plane strain and model x-y plane. II-When you select a structural module like plane stress or plane stress, then you have to put supports in your model to make it structurally stable otherwise your model is not stable and FEM programs can’t deal with it. III-You input µ =4*3.14*10^(-7) or µ =2000*4*3.14*10^(-7) in the B= µ0 µr H but 4*3.14*10^(-7) is not necessary. Because µr?is the relative permeability of the medium with respect to air’s permeability (µ0) so the minimum value for µr is 1 when the medium is air. ?0 = 4?×10?7 which is for the air exist already in the equation B= µ0 µr H. I try to look more in your model but can you fix the above issues and send it again Nastaran
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 15 sept. 2009, 06:58 UTC−4
Jigar,
I cannot help much but I give you some comments
The problem you are trying o solve can be easy or difficult. There are two scenarios
First:
You just want to know the forces acting on the iron particles and not the stress or displacement.
You can solve this problem with magnetostatic without coupling to structural mode. I attach a sample 3D file. Your model is very similar to the permanent magnet model in the AC/DC manual. In the file that I attached if you go to Data display/Global and you type P1_forcex_emnc you can we the force on the first particle along x direction and with P1_forcey_emnc and P1_forcez_emnc you can see the force value along y and z direction.
Second:
If you want to know the displacement of particles or the stress on the particles, then you need to couple magnetostatic to structural mode. But when you are in structural mode, you need to apply support on the Iron particles otherwise you cannot evaluate stress or displacement on the particles. But the problem is the complexity of the support in your model. The Iron particles are fixed in place because of the tissue around them if I am not wrong? So all over the particle is partially fixed and the particles are partially fixed in their place because of the contact with the tissue or the viscosity of the material around them. right? You may be able to apply some sort of spring on the particles to fix them in the domain partially .After you apply the supports on the particles then you have to couple magnetostatic and structural mode by for example applying P1_forcex_emnc and P1_forcey_emnc and P1_forcez_emnc as body forces to the particles in the structural mode.
Nastaran
Jigar, I cannot help much but I give you some comments The problem you are trying o solve can be easy or difficult. There are two scenarios First: You just want to know the forces acting on the iron particles and not the stress or displacement. You can solve this problem with magnetostatic without coupling to structural mode. I attach a sample 3D file. Your model is very similar to the permanent magnet model in the AC/DC manual. In the file that I attached if you go to Data display/Global and you type P1_forcex_emnc you can we the force on the first particle along x direction and with P1_forcey_emnc and P1_forcez_emnc you can see the force value along y and z direction. Second: If you want to know the displacement of particles or the stress on the particles, then you need to couple magnetostatic to structural mode. But when you are in structural mode, you need to apply support on the Iron particles otherwise you cannot evaluate stress or displacement on the particles. But the problem is the complexity of the support in your model. The Iron particles are fixed in place because of the tissue around them if I am not wrong? So all over the particle is partially fixed and the particles are partially fixed in their place because of the contact with the tissue or the viscosity of the material around them. right? You may be able to apply some sort of spring on the particles to fix them in the domain partially .After you apply the supports on the particles then you have to couple magnetostatic and structural mode by for example applying P1_forcex_emnc and P1_forcey_emnc and P1_forcez_emnc as body forces to the particles in the structural mode. Nastaran
Jigar Patel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 15 sept. 2009, 15:40 UTC−4

Jigar,
I cannot help much but I give you some comments
The problem you are trying o solve can be easy or difficult. There are two scenarios
First:
You just want to know the forces acting on the iron particles and not the stress or displacement.
You can solve this problem with magnetostatic without coupling to structural mode. I attach a sample 3D file. Your model is very similar to the permanent magnet model in the AC/DC manual. In the file that I attached if you go to Data display/Global and you type P1_forcex_emnc you can we the force on the first particle along x direction and with P1_forcey_emnc and P1_forcez_emnc you can see the force value along y and z direction.
Second:
If you want to know the displacement of particles or the stress on the particles, then you need to couple magnetostatic to structural mode. But when you are in structural mode, you need to apply support on the Iron particles otherwise you cannot evaluate stress or displacement on the particles. But the problem is the complexity of the support in your model. The Iron particles are fixed in place because of the tissue around them if I am not wrong? So all over the particle is partially fixed and the particles are partially fixed in their place because of the contact with the tissue or the viscosity of the material around them. right? You may be able to apply some sort of spring on the particles to fix them in the domain partially .After you apply the supports on the particles then you have to couple magnetostatic and structural mode by for example applying P1_forcex_emnc and P1_forcey_emnc and P1_forcez_emnc as body forces to the particles in the structural mode.
Nastaran



Nastaran,

I wasn't able to open up the file. However, I looked at your previous message, and deleted the 4*pi*10^(-7) values as they were unnecessary. You mentioned two cases that I could possibly be looking at, and I am interested in the second one. As you suggested, since I am conducting this problem first in 2D, I removed the plane strain model. You are right about the force acting on the particle. In this model I want to do a simple elastic force to prevent the particle from moving. In Matlab, this elastic force acting on the particle in x direction is written as K*(x0-x), where K is some constant, x0 is the initial location of the particle, and x is the position through time. I applied the plane stress model to both particles as well, and saw that there was a load section defined. However, I am not sure how to enter in that force. I also tried writing P1_forcex_emnc as you suggested in the load tab under subdomain settings, but when I try to run the simulation it says variables are not defined. I have attached the fixed file. Please let me know what you think.

Thanks,
Jigar
[QUOTE] Jigar, I cannot help much but I give you some comments The problem you are trying o solve can be easy or difficult. There are two scenarios First: You just want to know the forces acting on the iron particles and not the stress or displacement. You can solve this problem with magnetostatic without coupling to structural mode. I attach a sample 3D file. Your model is very similar to the permanent magnet model in the AC/DC manual. In the file that I attached if you go to Data display/Global and you type P1_forcex_emnc you can we the force on the first particle along x direction and with P1_forcey_emnc and P1_forcez_emnc you can see the force value along y and z direction. Second: If you want to know the displacement of particles or the stress on the particles, then you need to couple magnetostatic to structural mode. But when you are in structural mode, you need to apply support on the Iron particles otherwise you cannot evaluate stress or displacement on the particles. But the problem is the complexity of the support in your model. The Iron particles are fixed in place because of the tissue around them if I am not wrong? So all over the particle is partially fixed and the particles are partially fixed in their place because of the contact with the tissue or the viscosity of the material around them. right? You may be able to apply some sort of spring on the particles to fix them in the domain partially .After you apply the supports on the particles then you have to couple magnetostatic and structural mode by for example applying P1_forcex_emnc and P1_forcey_emnc and P1_forcez_emnc as body forces to the particles in the structural mode. Nastaran [/QUOTE] Nastaran, I wasn't able to open up the file. However, I looked at your previous message, and deleted the 4*pi*10^(-7) values as they were unnecessary. You mentioned two cases that I could possibly be looking at, and I am interested in the second one. As you suggested, since I am conducting this problem first in 2D, I removed the plane strain model. You are right about the force acting on the particle. In this model I want to do a simple elastic force to prevent the particle from moving. In Matlab, this elastic force acting on the particle in x direction is written as K*(x0-x), where K is some constant, x0 is the initial location of the particle, and x is the position through time. I applied the plane stress model to both particles as well, and saw that there was a load section defined. However, I am not sure how to enter in that force. I also tried writing P1_forcex_emnc as you suggested in the load tab under subdomain settings, but when I try to run the simulation it says variables are not defined. I have attached the fixed file. Please let me know what you think. Thanks, Jigar

Attachments:

Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 15 sept. 2009, 23:24 UTC−4
Jigar,
I made the model with the version 3.5 .There may be a version difference that you couldn’t open the attachment. I attach the file again if you couldn’t open it give me an email address so that I email it to you directly. This fie is only magnetostatic mode in 3D. I think it is better that first you make the magnetostatic mode work correctly and later couple it to structural mode
A few comments
1-I am not sure if using 2D model is the right solution because in 2D plane stress we actually model a cross section of the 3D problem with constant geometry along z direction but the particles in your model are sphere(and not cylinder) so it is not uniform along z direction. Actually the model I made is 3D solid stress strain coupled with magnetostatic-no current(emnc)
2-The reason that you see the error “variables are not defined”
First, you should be in 3D magnetostatic-no current(emnc) application mode
Then you go to subdomain settings in magnetostatic mode
In the “forces” tab for the subdomain number related to the particle 1 type P1 under the “Name” and P2 for particle 2 under the “Name”
Then run the model again
Then go to Data display/Global
In the expression type P1_forcex_emnc or P2_forcex_emnc. You should not see the error again
Sorry but I don’t know how to do it in the 2D
Nastaran

Jigar, I made the model with the version 3.5 .There may be a version difference that you couldn’t open the attachment. I attach the file again if you couldn’t open it give me an email address so that I email it to you directly. This fie is only magnetostatic mode in 3D. I think it is better that first you make the magnetostatic mode work correctly and later couple it to structural mode A few comments 1-I am not sure if using 2D model is the right solution because in 2D plane stress we actually model a cross section of the 3D problem with constant geometry along z direction but the particles in your model are sphere(and not cylinder) so it is not uniform along z direction. Actually the model I made is 3D solid stress strain coupled with magnetostatic-no current(emnc) 2-The reason that you see the error “variables are not defined” First, you should be in 3D magnetostatic-no current(emnc) application mode Then you go to subdomain settings in magnetostatic mode In the “forces” tab for the subdomain number related to the particle 1 type P1 under the “Name” and P2 for particle 2 under the “Name” Then run the model again Then go to Data display/Global In the expression type P1_forcex_emnc or P2_forcex_emnc. You should not see the error again Sorry but I don’t know how to do it in the 2D Nastaran
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 15 sept. 2009, 23:25 UTC−4
Jigar
The attachment didn't work
I try it again
Nastaran
Jigar The attachment didn't work I try it again Nastaran
Jigar Patel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 16 sept. 2009, 18:42 UTC−4

Jigar
The attachment didn't work
I try it again
Nastaran


Nastaran,

I actually figured out how to couple the force and now the particles are actually showing displacement and force. I had a question for you though. Is there a way to see the data of how a particle moved instead of a surface plot? I have been just clicking on the particle and seeing the starting and ending point but I was hoping that there was an actual way of doing this.

Thanks,
Jigar
[QUOTE] Jigar The attachment didn't work I try it again Nastaran [/QUOTE] Nastaran, I actually figured out how to couple the force and now the particles are actually showing displacement and force. I had a question for you though. Is there a way to see the data of how a particle moved instead of a surface plot? I have been just clicking on the particle and seeing the starting and ending point but I was hoping that there was an actual way of doing this. Thanks, Jigar
Jigar Patel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 16 sept. 2009, 18:52 UTC−4

Jigar
The attachment didn't work
I try it again
Nastaran


Nastaran,

Also I had some other questions. Do you know why the Comsol is not showing the particles moving up towards the magnet? I am assuming that the simulation is not running for long enough. How do I check the simulation time that it runs for? And is there a way to change that? I assumed changing from static to time harmonic but I don't think that is right.

Thanks,
Jigar
[QUOTE] Jigar The attachment didn't work I try it again Nastaran [/QUOTE] Nastaran, Also I had some other questions. Do you know why the Comsol is not showing the particles moving up towards the magnet? I am assuming that the simulation is not running for long enough. How do I check the simulation time that it runs for? And is there a way to change that? I assumed changing from static to time harmonic but I don't think that is right. Thanks, Jigar
Nastaran Tamjidi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 17 sept. 2009, 03:09 UTC−4
Jigar,
It is not the problem of COMSOL.
You are doing static analysis and the definition of static analysis is that the displacement doesn't depend on time. In static analysis we say that for example if we apply a 10N load at the end of a bar for 1 second or 10 second the displacement of the bar is the same so increasing the time doesn't change the final displacemnet.
In static analysis you can not change the running time because equations are independent of time.If you have time harmonic or time transient analysis then you can increase the time.But in your problem you don't need transient analysis because the force from permanant magnet to the particle is constant and independant of time.But if you just want to see a movie of particles movement you can try plot parameters/animation.
Nastaran
Jigar, It is not the problem of COMSOL. You are doing static analysis and the definition of static analysis is that the displacement doesn't depend on time. In static analysis we say that for example if we apply a 10N load at the end of a bar for 1 second or 10 second the displacement of the bar is the same so increasing the time doesn't change the final displacemnet. In static analysis you can not change the running time because equations are independent of time.If you have time harmonic or time transient analysis then you can increase the time.But in your problem you don't need transient analysis because the force from permanant magnet to the particle is constant and independant of time.But if you just want to see a movie of particles movement you can try plot parameters/animation. Nastaran
Jigar Patel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 24 sept. 2009, 13:35 UTC−4

Jigar,
It is not the problem of COMSOL.
You are doing static analysis and the definition of static analysis is that the displacement doesn't depend on time. In static analysis we say that for example if we apply a 10N load at the end of a bar for 1 second or 10 second the displacement of the bar is the same so increasing the time doesn't change the final displacemnet.
In static analysis you can not change the running time because equations are independent of time.If you have time harmonic or time transient analysis then you can increase the time.But in your problem you don't need transient analysis because the force from permanant magnet to the particle is constant and independant of time.But if you just want to see a movie of particles movement you can try plot parameters/animation.
Nastaran


Nastaran,

I see what you're saying, but I am thinking that my problem might be too complex for Comsol to solve. The issue is I have a code generated in Matlab with all the forces, and when I enter in specific locations, I can see that the particle moves towards the magnet. In Comsol, it shows that there is displacement, but the particle never moves towards the magnet. At first I figured I was entering the forces wrong, but I rechecked and they seem quite accurate. It doesn't make any sense to me why the particles won't move up towards the magnet no matter how close I put the particles. I tried the time transient also but no luck on that either. Realistically if I place a particle 0.5 mm away from a strong magnet, that particle should move towards it. However, Comsol only shows displacements in the order of 10^(-15).

Thanks,
Jigar
[QUOTE] Jigar, It is not the problem of COMSOL. You are doing static analysis and the definition of static analysis is that the displacement doesn't depend on time. In static analysis we say that for example if we apply a 10N load at the end of a bar for 1 second or 10 second the displacement of the bar is the same so increasing the time doesn't change the final displacemnet. In static analysis you can not change the running time because equations are independent of time.If you have time harmonic or time transient analysis then you can increase the time.But in your problem you don't need transient analysis because the force from permanant magnet to the particle is constant and independant of time.But if you just want to see a movie of particles movement you can try plot parameters/animation. Nastaran [/QUOTE] Nastaran, I see what you're saying, but I am thinking that my problem might be too complex for Comsol to solve. The issue is I have a code generated in Matlab with all the forces, and when I enter in specific locations, I can see that the particle moves towards the magnet. In Comsol, it shows that there is displacement, but the particle never moves towards the magnet. At first I figured I was entering the forces wrong, but I rechecked and they seem quite accurate. It doesn't make any sense to me why the particles won't move up towards the magnet no matter how close I put the particles. I tried the time transient also but no luck on that either. Realistically if I place a particle 0.5 mm away from a strong magnet, that particle should move towards it. However, Comsol only shows displacements in the order of 10^(-15). Thanks, Jigar
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 26 sept. 2009, 17:58 UTC−4
Hi

Have you guys taken a look at the model gallery, i.e.

www.comsol.com/showroom/documentation/model/197/

I beleive you can find some interesting ideas here and on a few similar cases

Ivar
Hi Have you guys taken a look at the model gallery, i.e. http://www.comsol.com/showroom/documentation/model/197/ I beleive you can find some interesting ideas here and on a few similar cases 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.

Suggested Content