Menu

Discussion Forum

New Discussion
查找

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 The information provided may be out of date.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.

how to calculate the mutual inductance of two coils

Posted 2011年3月19日 GMT-4 20:24 Version 4.0, Version 4.1 23 Replies

Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Dear all,

I am a beginner of COMSOL and I begin to try to use the 4.0 version several days ago.

Now I have the problem to calculate the mutual inductance of two or more coils which are placed into the media with electric conductivity range from 0.0005 S/m to 1000 S/m.

I created a simple model but don't know how to extract the inductance matrix of the coils system. Furthermore, there are still some problems to run the model. Could anyone help me to modify the model to let it output a table of inductance matrix when the electric conductivity changes?


Thanks,
Zhanghong Tang

Attachments:

23 Replies Last Post 2011年4月13日 GMT-4 12:16
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年3月19日 GMT-4 20:38

Dear all,

I am a beginner of COMSOL and I begin to try to use the 4.0 version several days ago.

Now I have the problem to calculate the mutual inductance of two or more coils which are placed into the media with electric conductivity range from 0.0005 S/m to 1000 S/m.

I created a simple model but don't know how to extract the inductance matrix of the coils system. Furthermore, there are still some problems to run the model. Could anyone help me to modify the model to let it output a table of inductance matrix when the electric conductivity changes?


Thanks,
Zhanghong Tang


Forgot to say, the frequency is about 20 kHz.

The original problem is to calculate the Induced voltage of receive coil when transmit coil have a current souce with frequency 20 kHz. The coils have hundreds turns which can't be acted a point, so that if we calculate the induced voltage by integrating the electrical field along the circle, we should calculate the integration several hundred times. One paper

“Prediction and Use of Impedance Matrices for Eddy-Current Problems,” IEEE Transactions on Magnetics, Kent R. Davey and Dalian Zheng, vol. 33 pp. 2478-2485, 1997.

said that we can calculate the impedance matrix of coils system at first, and then the induced voltage can be calculated very easy. So I want to calculate the impedance matrix of coils system.


Thanks,
Zhanghong Tang
[QUOTE] Dear all, I am a beginner of COMSOL and I begin to try to use the 4.0 version several days ago. Now I have the problem to calculate the mutual inductance of two or more coils which are placed into the media with electric conductivity range from 0.0005 S/m to 1000 S/m. I created a simple model but don't know how to extract the inductance matrix of the coils system. Furthermore, there are still some problems to run the model. Could anyone help me to modify the model to let it output a table of inductance matrix when the electric conductivity changes? Thanks, Zhanghong Tang [/QUOTE] Forgot to say, the frequency is about 20 kHz. The original problem is to calculate the Induced voltage of receive coil when transmit coil have a current souce with frequency 20 kHz. The coils have hundreds turns which can't be acted a point, so that if we calculate the induced voltage by integrating the electrical field along the circle, we should calculate the integration several hundred times. One paper “Prediction and Use of Impedance Matrices for Eddy-Current Problems,” IEEE Transactions on Magnetics, Kent R. Davey and Dalian Zheng, vol. 33 pp. 2478-2485, 1997. said that we can calculate the impedance matrix of coils system at first, and then the induced voltage can be calculated very easy. So I want to calculate the impedance matrix of coils system. Thanks, Zhanghong Tang
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: 1 decade ago 2011年3月20日 GMT-4 05:04
Hi

take a closer look at the lumped port BC's and the theory behind, well described in the latest release (V4.1 and better to come 4.2)

--
Good luck
Ivar
Hi take a closer look at the lumped port BC's and the theory behind, well described in the latest release (V4.1 and better to come 4.2) -- Good luck Ivar
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年3月20日 GMT-4 11:14

Hi

take a closer look at the lumped port BC's and the theory behind, well described in the latest release (V4.1 and better to come 4.2)

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply.
Do you mean that latest version V4.1 or upcoming V4.2 can calculate the impedance matrix directly?
On the other hand, this model I uploaded still have some problems when computation. Could you please help me to check the problem of this model and modify the model for me?

I wish I can have a solution of the field at the first step. I will try to integrate the electrical field to get the induced voltage before I can try the latest version.

Thanks,
Zhanghong Tang

[QUOTE] Hi take a closer look at the lumped port BC's and the theory behind, well described in the latest release (V4.1 and better to come 4.2) -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. Do you mean that latest version V4.1 or upcoming V4.2 can calculate the impedance matrix directly? On the other hand, this model I uploaded still have some problems when computation. Could you please help me to check the problem of this model and modify the model for me? I wish I can have a solution of the field at the first step. I will try to integrate the electrical field to get the induced voltage before I can try the latest version. Thanks, Zhanghong Tang
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: 1 decade ago 2011年3月20日 GMT-4 15:27
Hi

I believe its should be possible also in v4.0, but the doc was not fully ready then. I'll take a quick look again on your model. OK in 3D we do not have the lumped port approach

A few comments:

your infinite domain should be a "shell" and not the full volume. Personally I prefer to use a sphere and you need to set the PML type to cylindrical (in your case) or spherical for a sphere

It would be easier to check your model if you respected the units ;)

Your model is really axi-symmetric, which means that a 2D-axi could also be used.

But, I agree, I'm (partly) wrong there, the lumped ports are in the ES modules and not easy to find in MF/MEF (apart under the magnetic isolation node, but then I'm also lost w.r.t. linking to a coil current, I would need to dig into a ACDC course, but I do not have any for the new v4) ;)

--
Good luck
Ivar
Hi I believe its should be possible also in v4.0, but the doc was not fully ready then. I'll take a quick look again on your model. OK in 3D we do not have the lumped port approach A few comments: your infinite domain should be a "shell" and not the full volume. Personally I prefer to use a sphere and you need to set the PML type to cylindrical (in your case) or spherical for a sphere It would be easier to check your model if you respected the units ;) Your model is really axi-symmetric, which means that a 2D-axi could also be used. But, I agree, I'm (partly) wrong there, the lumped ports are in the ES modules and not easy to find in MF/MEF (apart under the magnetic isolation node, but then I'm also lost w.r.t. linking to a coil current, I would need to dig into a ACDC course, but I do not have any for the new v4) ;) -- Good luck Ivar
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年3月20日 GMT-4 19:26

Hi

I believe its should be possible also in v4.0, but the doc was not fully ready then. I'll take a quick look again on your model. OK in 3D we do not have the lumped port approach

A few comments:

your infinite domain should be a "shell" and not the full volume. Personally I prefer to use a sphere and you need to set the PML type to cylindrical (in your case) or spherical for a sphere

It would be easier to check your model if you respected the units ;)

Your model is really axi-symmetric, which means that a 2D-axi could also be used.

But, I agree, I'm (partly) wrong there, the lumped ports are in the ES modules and not easy to find in MF/MEF (apart under the magnetic isolation node, but then I'm also lost w.r.t. linking to a coil current, I would need to dig into a ACDC course, but I do not have any for the new v4) ;)

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your so kindly reply and your good suggestions. I will check the model and try to do some modification as you suggested.

1) How to set the "Shell" for infinite domain? Select the domains faces and then assign to the infinite element? The real dimension of the model could be: distances of coils are several meters, but the diameters of coils are only tens of millimeters. Is it still suitable to use sphere?
2) The default unit is meter, isn't it?
3) The real model is much more complex and it should be 3D. Currently I just built a sample model to check whether it works and whether it agrees with theorial result.

I am expecting to your modified model to let me better understand how to use COMSOL in our projects.


Thanks,
Zhanghong Tang
[QUOTE] Hi I believe its should be possible also in v4.0, but the doc was not fully ready then. I'll take a quick look again on your model. OK in 3D we do not have the lumped port approach A few comments: your infinite domain should be a "shell" and not the full volume. Personally I prefer to use a sphere and you need to set the PML type to cylindrical (in your case) or spherical for a sphere It would be easier to check your model if you respected the units ;) Your model is really axi-symmetric, which means that a 2D-axi could also be used. But, I agree, I'm (partly) wrong there, the lumped ports are in the ES modules and not easy to find in MF/MEF (apart under the magnetic isolation node, but then I'm also lost w.r.t. linking to a coil current, I would need to dig into a ACDC course, but I do not have any for the new v4) ;) -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your so kindly reply and your good suggestions. I will check the model and try to do some modification as you suggested. 1) How to set the "Shell" for infinite domain? Select the domains faces and then assign to the infinite element? The real dimension of the model could be: distances of coils are several meters, but the diameters of coils are only tens of millimeters. Is it still suitable to use sphere? 2) The default unit is meter, isn't it? 3) The real model is much more complex and it should be 3D. Currently I just built a sample model to check whether it works and whether it agrees with theorial result. I am expecting to your modified model to let me better understand how to use COMSOL in our projects. Thanks, Zhanghong Tang
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: 1 decade ago 2011年3月21日 GMT-4 02:01
Hi

it is perfectly possible to use a cylinder, but the way a magnetic field fall off with distance should also be taken into account. So check that the shape is not influencing the field lines more than expected, apply your critical "physics" sens to the results. Anyhow infinite elements or PML are just a shell, not the full volume

the default units are SI, and "m" for the geometry, except if you overwrite this in the units settings or in the geometry

Well 3D is fine, the only thing is that it takes time to solve, hence to debug. Then you are also approaching the issue the correct way: start simple, check the results where analytical approaches might be used, then complexify.

Mutual inductance is a classical case, so there must be others out here that has done the exercise

--
Good luck
Ivar
Hi it is perfectly possible to use a cylinder, but the way a magnetic field fall off with distance should also be taken into account. So check that the shape is not influencing the field lines more than expected, apply your critical "physics" sens to the results. Anyhow infinite elements or PML are just a shell, not the full volume the default units are SI, and "m" for the geometry, except if you overwrite this in the units settings or in the geometry Well 3D is fine, the only thing is that it takes time to solve, hence to debug. Then you are also approaching the issue the correct way: start simple, check the results where analytical approaches might be used, then complexify. Mutual inductance is a classical case, so there must be others out here that has done the exercise -- Good luck Ivar
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年3月30日 GMT-4 23:07

Hi

take a closer look at the lumped port BC's and the theory behind, well described in the latest release (V4.1 and better to come 4.2)

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply. I have the chance to read the document of latest version 4.1 and it did say that the lumped parameters can be calculated. It seems that COMSOL can get the induced voltage of coils directly so I don't need to calculate the inductance matrix at all.

Now my problem is: how to set parameters to let the COMSOL output the induced voltage of receive coils when transmit coils excite a current with frequency 20 kHz? Furthermore, when the conductivity of infinite uniform medium change from 0.0005 S/m to 1000 S/m, how can I output the induced voltage change with the conductivity (the sweep parameter is conductivity)?

I tested some other software such as Ansoft Maxwell and found that it can't calculate it when the conductivity is less than 1 S/m, I don't know whether the COMSOL has the same problem or not.

Thanks,
Zhanghong Tang
[QUOTE] Hi take a closer look at the lumped port BC's and the theory behind, well described in the latest release (V4.1 and better to come 4.2) -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. I have the chance to read the document of latest version 4.1 and it did say that the lumped parameters can be calculated. It seems that COMSOL can get the induced voltage of coils directly so I don't need to calculate the inductance matrix at all. Now my problem is: how to set parameters to let the COMSOL output the induced voltage of receive coils when transmit coils excite a current with frequency 20 kHz? Furthermore, when the conductivity of infinite uniform medium change from 0.0005 S/m to 1000 S/m, how can I output the induced voltage change with the conductivity (the sweep parameter is conductivity)? I tested some other software such as Ansoft Maxwell and found that it can't calculate it when the conductivity is less than 1 S/m, I don't know whether the COMSOL has the same problem or not. Thanks, Zhanghong Tang
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年3月31日 GMT-4 08:36


Hi

take a closer look at the lumped port BC's and the theory behind, well described in the latest release (V4.1 and better to come 4.2)

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply. I have the chance to read the document of latest version 4.1 and it did say that the lumped parameters can be calculated. It seems that COMSOL can get the induced voltage of coils directly so I don't need to calculate the inductance matrix at all.

Now my problem is: how to set parameters to let the COMSOL output the induced voltage of receive coils when transmit coils excite a current with frequency 20 kHz? Furthermore, when the conductivity of infinite uniform medium change from 0.0005 S/m to 1000 S/m, how can I output the induced voltage change with the conductivity (the sweep parameter is conductivity)?

I tested some other software such as Ansoft Maxwell and found that it can't calculate it when the conductivity is less than 1 S/m, I don't know whether the COMSOL has the same problem or not.

Thanks,
Zhanghong Tang


Dear Dr. Ivar,

In page 139 of the document "ACDCModuleUserGuide.pdf", it pointed out that the variables such as impedance, resistance and so on are stored in the global variables. However, it is only for 2D axisymmetric models. What's the variables defined for general 3D models?

Thanks,
Zhanghong Tang
[QUOTE] [QUOTE] Hi take a closer look at the lumped port BC's and the theory behind, well described in the latest release (V4.1 and better to come 4.2) -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. I have the chance to read the document of latest version 4.1 and it did say that the lumped parameters can be calculated. It seems that COMSOL can get the induced voltage of coils directly so I don't need to calculate the inductance matrix at all. Now my problem is: how to set parameters to let the COMSOL output the induced voltage of receive coils when transmit coils excite a current with frequency 20 kHz? Furthermore, when the conductivity of infinite uniform medium change from 0.0005 S/m to 1000 S/m, how can I output the induced voltage change with the conductivity (the sweep parameter is conductivity)? I tested some other software such as Ansoft Maxwell and found that it can't calculate it when the conductivity is less than 1 S/m, I don't know whether the COMSOL has the same problem or not. Thanks, Zhanghong Tang [/QUOTE] Dear Dr. Ivar, In page 139 of the document "ACDCModuleUserGuide.pdf", it pointed out that the variables such as impedance, resistance and so on are stored in the global variables. However, it is only for 2D axisymmetric models. What's the variables defined for general 3D models? Thanks, Zhanghong Tang
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: 1 decade ago 2011年3月31日 GMT-4 09:49
Hi

the change in conductivity and solver issues might come from a too large numerical difference (dependent variable scaling scaling issue, you might need to do some manual scaling). In principle I do not see why it could not work even for 1mS/m

To plot the results (in v4) versus the parameter sweep use a table filled in by a "derived varibale" over the study set

--
Good luck
Ivar
Hi the change in conductivity and solver issues might come from a too large numerical difference (dependent variable scaling scaling issue, you might need to do some manual scaling). In principle I do not see why it could not work even for 1mS/m To plot the results (in v4) versus the parameter sweep use a table filled in by a "derived varibale" over the study set -- Good luck Ivar
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年3月31日 GMT-4 12:23

Hi

the change in conductivity and solver issues might come from a too large numerical difference (dependent variable scaling scaling issue, you might need to do some manual scaling). In principle I do not see why it could not work even for 1mS/m

To plot the results (in v4) versus the parameter sweep use a table filled in by a "derived varibale" over the study set

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply.
1) Could you please explain more about "do some manual scaling"?
2) Do you mean that even in 3D, I can also get the impedance value? What's the name of the variable (from the document I can get to know it is mf.Z_cg<name> for 2D axisymmetric modls, what's the name of the variable when the models are general 3D)?

Thanks,
ZHanghong Tang
[QUOTE] Hi the change in conductivity and solver issues might come from a too large numerical difference (dependent variable scaling scaling issue, you might need to do some manual scaling). In principle I do not see why it could not work even for 1mS/m To plot the results (in v4) versus the parameter sweep use a table filled in by a "derived varibale" over the study set -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. 1) Could you please explain more about "do some manual scaling"? 2) Do you mean that even in 3D, I can also get the impedance value? What's the name of the variable (from the document I can get to know it is mf.Z_cg for 2D axisymmetric modls, what's the name of the variable when the models are general 3D)? Thanks, ZHanghong Tang
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月1日 GMT-4 04:37


Hi

the change in conductivity and solver issues might come from a too large numerical difference (dependent variable scaling scaling issue, you might need to do some manual scaling). In principle I do not see why it could not work even for 1mS/m

To plot the results (in v4) versus the parameter sweep use a table filled in by a "derived varibale" over the study set

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply.
1) Could you please explain more about "do some manual scaling"?
2) Do you mean that even in 3D, I can also get the impedance value? What's the name of the variable (from the document I can get to know it is mf.Z_cg<name> for 2D axisymmetric modls, what's the name of the variable when the models are general 3D)?

Thanks,
ZHanghong Tang


Dear all,

I tried to build a 2D axisymmetric model by 4.1. However, it still can't work. The following errors displayed:
Error in sweep: The following feature has encountered a problem
-Feature: Stationary Solver 1 (sol1/s1)
-Error: Empty parameter list.

I set a parameter 'xigma' and set it to be sweep parameter in 'Study 1'. The 'xigma' is set as:
range(0.0005,10,1000)

I also don't know how to let the COMSOL plot the curve with x axis be 'xigma' and y axis be 'mf.Rcoil_cg<name>' as described in page 139 of 'ACDCModuleUserGuide.pdf'.

Could anyone help me to check the problems of the model, and tell me how to add this plot?

Is there any method to get the same result when the model is modified to 3D?

Thanks,
Zhanghong Tang

[QUOTE] [QUOTE] Hi the change in conductivity and solver issues might come from a too large numerical difference (dependent variable scaling scaling issue, you might need to do some manual scaling). In principle I do not see why it could not work even for 1mS/m To plot the results (in v4) versus the parameter sweep use a table filled in by a "derived varibale" over the study set -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. 1) Could you please explain more about "do some manual scaling"? 2) Do you mean that even in 3D, I can also get the impedance value? What's the name of the variable (from the document I can get to know it is mf.Z_cg for 2D axisymmetric modls, what's the name of the variable when the models are general 3D)? Thanks, ZHanghong Tang [/QUOTE] Dear all, I tried to build a 2D axisymmetric model by 4.1. However, it still can't work. The following errors displayed: Error in sweep: The following feature has encountered a problem -Feature: Stationary Solver 1 (sol1/s1) -Error: Empty parameter list. I set a parameter 'xigma' and set it to be sweep parameter in 'Study 1'. The 'xigma' is set as: range(0.0005,10,1000) I also don't know how to let the COMSOL plot the curve with x axis be 'xigma' and y axis be 'mf.Rcoil_cg' as described in page 139 of 'ACDCModuleUserGuide.pdf'. Could anyone help me to check the problems of the model, and tell me how to add this plot? Is there any method to get the same result when the model is modified to 3D? Thanks, Zhanghong Tang

Attachments:

Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月1日 GMT-4 08:32



Hi

the change in conductivity and solver issues might come from a too large numerical difference (dependent variable scaling scaling issue, you might need to do some manual scaling). In principle I do not see why it could not work even for 1mS/m

To plot the results (in v4) versus the parameter sweep use a table filled in by a "derived varibale" over the study set

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply.
1) Could you please explain more about "do some manual scaling"?
2) Do you mean that even in 3D, I can also get the impedance value? What's the name of the variable (from the document I can get to know it is mf.Z_cg<name> for 2D axisymmetric modls, what's the name of the variable when the models are general 3D)?

Thanks,
ZHanghong Tang


Dear all,

I tried to build a 2D axisymmetric model by 4.1. However, it still can't work. The following errors displayed:
Error in sweep: The following feature has encountered a problem
-Feature: Stationary Solver 1 (sol1/s1)
-Error: Empty parameter list.

I set a parameter 'xigma' and set it to be sweep parameter in 'Study 1'. The 'xigma' is set as:
range(0.0005,10,1000)

I also don't know how to let the COMSOL plot the curve with x axis be 'xigma' and y axis be 'mf.Rcoil_cg<name>' as described in page 139 of 'ACDCModuleUserGuide.pdf'.

Could anyone help me to check the problems of the model, and tell me how to add this plot?

Is there any method to get the same result when the model is modified to 3D?

Thanks,
Zhanghong Tang



Dear all,

I forgot to add the frequency in the attached model. After added that the simulation is done but at last, the following error information displayed:

Error in sweep: Java heap space


What should I do?

Thanks,
Zhanghong Tang
[QUOTE] [QUOTE] [QUOTE] Hi the change in conductivity and solver issues might come from a too large numerical difference (dependent variable scaling scaling issue, you might need to do some manual scaling). In principle I do not see why it could not work even for 1mS/m To plot the results (in v4) versus the parameter sweep use a table filled in by a "derived varibale" over the study set -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. 1) Could you please explain more about "do some manual scaling"? 2) Do you mean that even in 3D, I can also get the impedance value? What's the name of the variable (from the document I can get to know it is mf.Z_cg for 2D axisymmetric modls, what's the name of the variable when the models are general 3D)? Thanks, ZHanghong Tang [/QUOTE] Dear all, I tried to build a 2D axisymmetric model by 4.1. However, it still can't work. The following errors displayed: Error in sweep: The following feature has encountered a problem -Feature: Stationary Solver 1 (sol1/s1) -Error: Empty parameter list. I set a parameter 'xigma' and set it to be sweep parameter in 'Study 1'. The 'xigma' is set as: range(0.0005,10,1000) I also don't know how to let the COMSOL plot the curve with x axis be 'xigma' and y axis be 'mf.Rcoil_cg' as described in page 139 of 'ACDCModuleUserGuide.pdf'. Could anyone help me to check the problems of the model, and tell me how to add this plot? Is there any method to get the same result when the model is modified to 3D? Thanks, Zhanghong Tang [/QUOTE] Dear all, I forgot to add the frequency in the attached model. After added that the simulation is done but at last, the following error information displayed: Error in sweep: Java heap space What should I do? Thanks, Zhanghong Tang
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: 1 decade ago 2011年4月2日 GMT-4 10:41
Hi

check on the KB there are a few suggestions how to best increase the Java settings, hopefully it would solve your issue

--
Good luck
Ivar
Hi check on the KB there are a few suggestions how to best increase the Java settings, hopefully it would solve your issue -- Good luck Ivar
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月3日 GMT-4 09:00

Hi

check on the KB there are a few suggestions how to best increase the Java settings, hopefully it would solve your issue

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply.

Could you please give me more details about "how to best increase the Java settings"? I searched related articles from the website but have not succeed.

In addition, after the computation finished, how can I plot the curve of "xigma-mf.Rcoil_cg<name>" ?


Thanks,
Zhanghong Tang
[QUOTE] Hi check on the KB there are a few suggestions how to best increase the Java settings, hopefully it would solve your issue -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. Could you please give me more details about "how to best increase the Java settings"? I searched related articles from the website but have not succeed. In addition, after the computation finished, how can I plot the curve of "xigma-mf.Rcoil_cg" ? Thanks, Zhanghong Tang
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月3日 GMT-4 12:34


Hi

check on the KB there are a few suggestions how to best increase the Java settings, hopefully it would solve your issue

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply.

Could you please give me more details about "how to best increase the Java settings"? I searched related articles from the website but have not succeed.

In addition, after the computation finished, how can I plot the curve of "xigma-mf.Rcoil_cg<name>" ?


Thanks,
Zhanghong Tang



Dear all,

The Java problem is solver when referred to this page:
www.comsol.nl/community/forums/general/thread/6383/

However, I still don't know how to plot the curve of "xigma-mf.Rcoil_cg<name>" , could anyone give me some hints?


Thanks,
Zhanghong Tang
[QUOTE] [QUOTE] Hi check on the KB there are a few suggestions how to best increase the Java settings, hopefully it would solve your issue -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. Could you please give me more details about "how to best increase the Java settings"? I searched related articles from the website but have not succeed. In addition, after the computation finished, how can I plot the curve of "xigma-mf.Rcoil_cg" ? Thanks, Zhanghong Tang [/QUOTE] Dear all, The Java problem is solver when referred to this page: http://www.comsol.nl/community/forums/general/thread/6383/ However, I still don't know how to plot the curve of "xigma-mf.Rcoil_cg" , could anyone give me some hints? Thanks, Zhanghong Tang
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月6日 GMT-4 11:14



Hi

check on the KB there are a few suggestions how to best increase the Java settings, hopefully it would solve your issue

--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply.

Could you please give me more details about "how to best increase the Java settings"? I searched related articles from the website but have not succeed.

In addition, after the computation finished, how can I plot the curve of "xigma-mf.Rcoil_cg<name>" ?


Thanks,
Zhanghong Tang



Dear all,

The Java problem is solver when referred to this page:
www.comsol.nl/community/forums/general/thread/6383/

However, I still don't know how to plot the curve of "xigma-mf.Rcoil_cg<name>" , could anyone give me some hints?


Thanks,
Zhanghong Tang


Dear Dr. Ivar,

Could you please give me more detailed tips to plot the curve of "xigma" and "mf.Rcoil_cg<name>" (the relationship of xigma and real part of impedance)?

Thanks,
Zhangong Tang
[QUOTE] [QUOTE] [QUOTE] Hi check on the KB there are a few suggestions how to best increase the Java settings, hopefully it would solve your issue -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. Could you please give me more details about "how to best increase the Java settings"? I searched related articles from the website but have not succeed. In addition, after the computation finished, how can I plot the curve of "xigma-mf.Rcoil_cg" ? Thanks, Zhanghong Tang [/QUOTE] Dear all, The Java problem is solver when referred to this page: http://www.comsol.nl/community/forums/general/thread/6383/ However, I still don't know how to plot the curve of "xigma-mf.Rcoil_cg" , could anyone give me some hints? Thanks, Zhanghong Tang [/QUOTE] Dear Dr. Ivar, Could you please give me more detailed tips to plot the curve of "xigma" and "mf.Rcoil_cg" (the relationship of xigma and real part of impedance)? Thanks, Zhangong Tang
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: 1 decade ago 2011年4月7日 GMT-4 03:03
Hi

with this long thread and the recalls I'm getting lost. But when I look at the coil tests.mph file, there are a few things I do not understand:

1) the resistivity of the surrounding media, in MF will only change mf.Jiphi (Induced phi current), and locally Qrh
2) if you do a frequency domain sweep, you must define at least one frequency (I tried 50Hz)

Pls remind us what exactly you are looking for and latest update, otherwise I'm afraid we are lost here :)


--
Good luck
Ivar
Hi with this long thread and the recalls I'm getting lost. But when I look at the coil tests.mph file, there are a few things I do not understand: 1) the resistivity of the surrounding media, in MF will only change mf.Jiphi (Induced phi current), and locally Qrh 2) if you do a frequency domain sweep, you must define at least one frequency (I tried 50Hz) Pls remind us what exactly you are looking for and latest update, otherwise I'm afraid we are lost here :) -- Good luck Ivar
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月7日 GMT-4 05:09

Hi

with this long thread and the recalls I'm getting lost. But when I look at the coil tests.mph file, there are a few things I do not understand:

1) the resistivity of the surrounding media, in MF will only change mf.Jiphi (Induced phi current), and locally Qrh
2) if you do a frequency domain sweep, you must define at least one frequency (I tried 50Hz)

Pls remind us what exactly you are looking for and latest update, otherwise I'm afraid we are lost here :)


--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply. The latest file is "coiltest_2d_4.1.mph" and now the paremeter "xigma" can be sweeped as I wanted.

However, I don't know how to plot the curve of "xigma-mf.Rcoil_cg<name>" that I got from page 139 of the document "ACDCModuleUserGuide.pdf", the document said that the impedance matrix of coils can be got from the global variable.

Could you please give me more detailed tips to plot the curve of "xigma" and "mf.Rcoil_cg<name>" (the relationship of xigma and real part of impedance)?


Thanks,
Zhanghong Tang
[QUOTE] Hi with this long thread and the recalls I'm getting lost. But when I look at the coil tests.mph file, there are a few things I do not understand: 1) the resistivity of the surrounding media, in MF will only change mf.Jiphi (Induced phi current), and locally Qrh 2) if you do a frequency domain sweep, you must define at least one frequency (I tried 50Hz) Pls remind us what exactly you are looking for and latest update, otherwise I'm afraid we are lost here :) -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. The latest file is "coiltest_2d_4.1.mph" and now the paremeter "xigma" can be sweeped as I wanted. However, I don't know how to plot the curve of "xigma-mf.Rcoil_cg" that I got from page 139 of the document "ACDCModuleUserGuide.pdf", the document said that the impedance matrix of coils can be got from the global variable. Could you please give me more detailed tips to plot the curve of "xigma" and "mf.Rcoil_cg" (the relationship of xigma and real part of impedance)? Thanks, Zhanghong Tang

Attachments:

Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月7日 GMT-4 21:30


Hi

with this long thread and the recalls I'm getting lost. But when I look at the coil tests.mph file, there are a few things I do not understand:

1) the resistivity of the surrounding media, in MF will only change mf.Jiphi (Induced phi current), and locally Qrh
2) if you do a frequency domain sweep, you must define at least one frequency (I tried 50Hz)

Pls remind us what exactly you are looking for and latest update, otherwise I'm afraid we are lost here :)


--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply. The latest file is "coiltest_2d_4.1.mph" and now the paremeter "xigma" can be sweeped as I wanted.

However, I don't know how to plot the curve of "xigma-mf.Rcoil_cg<name>" that I got from page 139 of the document "ACDCModuleUserGuide.pdf", the document said that the impedance matrix of coils can be got from the global variable.

Could you please give me more detailed tips to plot the curve of "xigma" and "mf.Rcoil_cg<name>" (the relationship of xigma and real part of impedance)?


Thanks,
Zhanghong Tang


Dear Dr. Ivar,

Have I missed anything to figure out my question?

Thanks,
Zhanghong Tang
[QUOTE] [QUOTE] Hi with this long thread and the recalls I'm getting lost. But when I look at the coil tests.mph file, there are a few things I do not understand: 1) the resistivity of the surrounding media, in MF will only change mf.Jiphi (Induced phi current), and locally Qrh 2) if you do a frequency domain sweep, you must define at least one frequency (I tried 50Hz) Pls remind us what exactly you are looking for and latest update, otherwise I'm afraid we are lost here :) -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. The latest file is "coiltest_2d_4.1.mph" and now the paremeter "xigma" can be sweeped as I wanted. However, I don't know how to plot the curve of "xigma-mf.Rcoil_cg" that I got from page 139 of the document "ACDCModuleUserGuide.pdf", the document said that the impedance matrix of coils can be got from the global variable. Could you please give me more detailed tips to plot the curve of "xigma" and "mf.Rcoil_cg" (the relationship of xigma and real part of impedance)? Thanks, Zhanghong Tang [/QUOTE] Dear Dr. Ivar, Have I missed anything to figure out my question? Thanks, Zhanghong Tang
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: 1 decade ago 2011年4月8日 GMT-4 02:07
Hi

I believe I start to understand, it's the name _cg that confuses me as it looks like a 3.5 name.

First of all that page is referring to "lumped parameter" calculations from a port or terminal, these exist when you solve for "V" the voltage hence in EC, ES, ... MEF but I do not believe in MF as V is not a dependent variable

But I'm no longer sure if MEF in 2D-axi is considering in plane current or out of plane currents for the loop current/voltage you are considering ?
ACDC is tricky as you have mostly 3 component fields interaction via the cross product, so its rather "limitative" to use 2D or 2D-axi representations, and one must clearly understand the hypothesis used

So I do not have any clear answer (I haven't been at the latest ACDC courses, would need to clarify this, as things are too different from 3.5 where these aspects were described in more details), sorry cannot do more here and now

One thing, it might work in MF with the new "single turn coil" domains BCs, but then you need to simulate your coil by a small circle and not a "point", try it out ;)

--
Good luck
Ivar
Hi I believe I start to understand, it's the name _cg that confuses me as it looks like a 3.5 name. First of all that page is referring to "lumped parameter" calculations from a port or terminal, these exist when you solve for "V" the voltage hence in EC, ES, ... MEF but I do not believe in MF as V is not a dependent variable But I'm no longer sure if MEF in 2D-axi is considering in plane current or out of plane currents for the loop current/voltage you are considering ? ACDC is tricky as you have mostly 3 component fields interaction via the cross product, so its rather "limitative" to use 2D or 2D-axi representations, and one must clearly understand the hypothesis used So I do not have any clear answer (I haven't been at the latest ACDC courses, would need to clarify this, as things are too different from 3.5 where these aspects were described in more details), sorry cannot do more here and now One thing, it might work in MF with the new "single turn coil" domains BCs, but then you need to simulate your coil by a small circle and not a "point", try it out ;) -- Good luck Ivar
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月8日 GMT-4 02:48



Hi

with this long thread and the recalls I'm getting lost. But when I look at the coil tests.mph file, there are a few things I do not understand:

1) the resistivity of the surrounding media, in MF will only change mf.Jiphi (Induced phi current), and locally Qrh
2) if you do a frequency domain sweep, you must define at least one frequency (I tried 50Hz)

Pls remind us what exactly you are looking for and latest update, otherwise I'm afraid we are lost here :)


--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply. The latest file is "coiltest_2d_4.1.mph" and now the paremeter "xigma" can be sweeped as I wanted.

However, I don't know how to plot the curve of "xigma-mf.Rcoil_cg<name>" that I got from page 139 of the document "ACDCModuleUserGuide.pdf", the document said that the impedance matrix of coils can be got from the global variable.

Could you please give me more detailed tips to plot the curve of "xigma" and "mf.Rcoil_cg<name>" (the relationship of xigma and real part of impedance)?


Thanks,
Zhanghong Tang


Dear Dr. Ivar,

Have I missed anything to figure out my question?

Thanks,
Zhanghong Tang


Dear Dr. Ivar,

My real problem is several coils with multiple turns. The sections of coils are rectangles. The current frequency of transmit coil is 20 kHz. I need to simulate the induced voltages of received coil when the "xigma" changed and plot the curve of "xigma-induced voltage". Could you please help me to create such an example model?

Thanks,
Zhanghong Tang
[QUOTE] [QUOTE] [QUOTE] Hi with this long thread and the recalls I'm getting lost. But when I look at the coil tests.mph file, there are a few things I do not understand: 1) the resistivity of the surrounding media, in MF will only change mf.Jiphi (Induced phi current), and locally Qrh 2) if you do a frequency domain sweep, you must define at least one frequency (I tried 50Hz) Pls remind us what exactly you are looking for and latest update, otherwise I'm afraid we are lost here :) -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. The latest file is "coiltest_2d_4.1.mph" and now the paremeter "xigma" can be sweeped as I wanted. However, I don't know how to plot the curve of "xigma-mf.Rcoil_cg" that I got from page 139 of the document "ACDCModuleUserGuide.pdf", the document said that the impedance matrix of coils can be got from the global variable. Could you please give me more detailed tips to plot the curve of "xigma" and "mf.Rcoil_cg" (the relationship of xigma and real part of impedance)? Thanks, Zhanghong Tang [/QUOTE] Dear Dr. Ivar, Have I missed anything to figure out my question? Thanks, Zhanghong Tang [/QUOTE] Dear Dr. Ivar, My real problem is several coils with multiple turns. The sections of coils are rectangles. The current frequency of transmit coil is 20 kHz. I need to simulate the induced voltages of received coil when the "xigma" changed and plot the curve of "xigma-induced voltage". Could you please help me to create such an example model? Thanks, Zhanghong Tang
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: 1 decade ago 2011年4月8日 GMT-4 03:03
Then I would say:

1) forget 4.0 but use latest release
2) use the new multi-turn coil BC's, but read carefully the doc
3) the point 2 means that you must design 2D areas that correspond to the true areas
4) use infinite elements to make the air somewhat smaller to not use too many elements
--
Good luck
Ivar
Then I would say: 1) forget 4.0 but use latest release 2) use the new multi-turn coil BC's, but read carefully the doc 3) the point 2 means that you must design 2D areas that correspond to the true areas 4) use infinite elements to make the air somewhat smaller to not use too many elements -- Good luck Ivar
Zhanghong Tang
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年4月13日 GMT-4 12:16

Then I would say:

1) forget 4.0 but use latest release
2) use the new multi-turn coil BC's, but read carefully the doc
3) the point 2 means that you must design 2D areas that correspond to the true areas
4) use infinite elements to make the air somewhat smaller to not use too many elements
--
Good luck
Ivar


Dear Dr. Ivar,

Thank you very much for your kindly reply. I have tried to installed 4.1, however, I still can't figure how to do as you described. Could you please give me an example model to show how to calculate the inducted voltage of the received coil when transmit coil has 1 A current with frequency 20 kHz, when the conductivity of infinite medium changes from 0.0005 S/m to 1000 S/m?

I wish to have a 2-D model at first, although my final model should be 3-D.

Thanks,
Zhanghong Tang
[QUOTE] Then I would say: 1) forget 4.0 but use latest release 2) use the new multi-turn coil BC's, but read carefully the doc 3) the point 2 means that you must design 2D areas that correspond to the true areas 4) use infinite elements to make the air somewhat smaller to not use too many elements -- Good luck Ivar [/QUOTE] Dear Dr. Ivar, Thank you very much for your kindly reply. I have tried to installed 4.1, however, I still can't figure how to do as you described. Could you please give me an example model to show how to calculate the inducted voltage of the received coil when transmit coil has 1 A current with frequency 20 kHz, when the conductivity of infinite medium changes from 0.0005 S/m to 1000 S/m? I wish to have a 2-D model at first, although my final model should be 3-D. Thanks, Zhanghong Tang

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