Menu

Discussion Forum

New Discussion
Filter

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.

Using saved solution Comsol v4.1

Posted Apr 1, 2011, 3:40 p.m. EDT MEMS & Nanotechnology, MEMS & Piezoelectric Devices, Structural Mechanics 14 Replies

Osameh Ghazian
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

hi,

I want to use the saved solution for the initialization of the second one. for example I solve my problem with a given boundary condition and I want to increase the boundary value for the next modeling. But for the new study I want to use the previous solved results as the initialization values of the new one. how can I do it ??

I have to press the Update Solution button ??

Thanks,
14 Replies Last Post Dec 25, 2015, 4:58 a.m. EST
Ruud Borger COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 2, 2011, 7:25 a.m. EDT
Dear Osameh,

The way to do this is to solve your initial model, then go back to Study 1. Open it, right-click "Solver 1" and press "Disable". This means that the solver sequence, and the solution attached to it, will not be changed. If you now right-click Study 1, and press "Show default solver", then a new solver sequence is built (Solver 2). Open this one, and in "Dependent variables", choose the solution from Solver 1 as initial values.

Tip #1: To keep an overview, it helps to rename "Solver 1" to "Solver V=3" or something related to your boundary condition, so that you can keep track of which solver contains what. The same goes for the solution.

Tip #2: You can automatize this process by using a parametric sweep.

Hope that helps,

Ruud Borger
COMSOLBV
Dear Osameh, The way to do this is to solve your initial model, then go back to Study 1. Open it, right-click "Solver 1" and press "Disable". This means that the solver sequence, and the solution attached to it, will not be changed. If you now right-click Study 1, and press "Show default solver", then a new solver sequence is built (Solver 2). Open this one, and in "Dependent variables", choose the solution from Solver 1 as initial values. Tip #1: To keep an overview, it helps to rename "Solver 1" to "Solver V=3" or something related to your boundary condition, so that you can keep track of which solver contains what. The same goes for the solution. Tip #2: You can automatize this process by using a parametric sweep. Hope that helps, Ruud Borger COMSOLBV
Osameh Ghazian
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 2, 2011, 9:15 a.m. EDT
Thank you so much

I have two sections named Dependent Variables in Solver 1 node, one of them is "Dependent variable 1" for transient initialization and the other one is 'Dependent variable 2" for time dependent.

1) In my new solver (Solver 2) I have to change the initial values of the both dependent variables sections ??

2) what is the "values not solved for" section ?

Best regards,



Thank you so much I have two sections named Dependent Variables in Solver 1 node, one of them is "Dependent variable 1" for transient initialization and the other one is 'Dependent variable 2" for time dependent. 1) In my new solver (Solver 2) I have to change the initial values of the both dependent variables sections ?? 2) what is the "values not solved for" section ? Best regards,
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 Apr 2, 2011, 10:37 a.m. EDT
Hi

when you have several physics you might select which ones are actives for each solver, hence you might want to add two solvers, sequently, and only solve one physics per soler node, then you need to decide what to do with the dependnet variables "not solved for"

--
Good luck
Ivar
Hi when you have several physics you might select which ones are actives for each solver, hence you might want to add two solvers, sequently, and only solve one physics per soler node, then you need to decide what to do with the dependnet variables "not solved for" -- Good luck Ivar
Ruud Borger COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 2, 2011, 12:55 p.m. EDT

Thank you so much

I have two sections named Dependent Variables in Solver 1 node, one of them is "Dependent variable 1" for transient initialization and the other one is 'Dependent variable 2" for time dependent.

1) In my new solver (Solver 2) I have to change the initial values of the both dependent variables sections ??

2) what is the "values not solved for" section ?

Best regards,


1) The transient initialization and time dependent studies are separate. The initialization probably does not need initial values (depends a bit on the physics interface(s) you are using). The time dependent part does benefit from them, use "Variables solved for" in this case.

2) "Variables not solved for" means exactly what it says. For instance, a fluid flow calculation may use the temperature from a different physics interface (Heat Transfer). In this case, the temperature T is a variable not solved for. So, if you first do a Heat Transfer calculation and then want to use these results in the fluid flow calculation, then you use the initial value under "Variables not solved for". If you want to use the results from a previous fluid flow calculation for a new flow calculation, then you use "Variables solved for".

Best regards,

Ruud Borger,
COMSOL
[QUOTE] Thank you so much I have two sections named Dependent Variables in Solver 1 node, one of them is "Dependent variable 1" for transient initialization and the other one is 'Dependent variable 2" for time dependent. 1) In my new solver (Solver 2) I have to change the initial values of the both dependent variables sections ?? 2) what is the "values not solved for" section ? Best regards, [/QUOTE] 1) The transient initialization and time dependent studies are separate. The initialization probably does not need initial values (depends a bit on the physics interface(s) you are using). The time dependent part does benefit from them, use "Variables solved for" in this case. 2) "Variables not solved for" means exactly what it says. For instance, a fluid flow calculation may use the temperature from a different physics interface (Heat Transfer). In this case, the temperature T is a variable not solved for. So, if you first do a Heat Transfer calculation and then want to use these results in the fluid flow calculation, then you use the initial value under "Variables not solved for". If you want to use the results from a previous fluid flow calculation for a new flow calculation, then you use "Variables solved for". Best regards, Ruud Borger, COMSOL
Sherif Abdelaziz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 3, 2011, 10:19 a.m. EDT
Would someone guide me about how to use parametric sweep to automate the changing boundary conditions case? if it is written some where in the documentations that accompanied COMSOL that will be great.

Thanks
Sherif
Would someone guide me about how to use parametric sweep to automate the changing boundary conditions case? if it is written some where in the documentations that accompanied COMSOL that will be great. Thanks Sherif
Alexis Grenier
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 3, 2011, 6:07 a.m. EDT
Hi,

I have a problem to use a saved solution.

The model has two physics : Heat Transer and Solid Mechanics.

I disable the solver 1 and create an other solver. I select the solution of solver 1 as initial values for solver 2 by the dependent variables node : Method = solution
Solution = Solver 1
Time = automatic

It is working well for the temperature. But the stresses of Solver 2 at t=0 are 0!

Any help would be greatly appreciated :)

Alexis



Hi, I have a problem to use a saved solution. The model has two physics : Heat Transer and Solid Mechanics. I disable the solver 1 and create an other solver. I select the solution of solver 1 as initial values for solver 2 by the dependent variables node : Method = solution Solution = Solver 1 Time = automatic It is working well for the temperature. But the stresses of Solver 2 at t=0 are 0! Any help would be greatly appreciated :) Alexis
Liu Xin
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 29, 2013, 2:57 a.m. EST

Would someone guide me about how to use parametric sweep to automate the changing boundary conditions case? if it is written some where in the documentations that accompanied COMSOL that will be great.

Thanks
Sherif


you can define a parameter under 'global definitions' and name it, for instant, 'para' with the value 0;
use the parameter in your changing boundary condition;
right click the study and add the 'parametric sweep', choose ' para' from the list, then you can define the range of the value to sweep.
[QUOTE] Would someone guide me about how to use parametric sweep to automate the changing boundary conditions case? if it is written some where in the documentations that accompanied COMSOL that will be great. Thanks Sherif [/QUOTE] you can define a parameter under 'global definitions' and name it, for instant, 'para' with the value 0; use the parameter in your changing boundary condition; right click the study and add the 'parametric sweep', choose ' para' from the list, then you can define the range of the value to sweep.
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 Jan 29, 2013, 1:56 p.m. EST
Hi

I have also issues to really understand when I get the initial conditions truely transferred, and when not,
I'm missing something like a line in the log file stating reading T from study 1 or something like that, and for each variable. I was woundering, if for the stress, if one shold add an initial stress sub node or not, I cannot find anything in the doc, which (sub)node is really populated: just the initial values, other subnodes ... ?

Here I find that a verbose log file really could help !

--
Good luck
Ivar
Hi I have also issues to really understand when I get the initial conditions truely transferred, and when not, I'm missing something like a line in the log file stating reading T from study 1 or something like that, and for each variable. I was woundering, if for the stress, if one shold add an initial stress sub node or not, I cannot find anything in the doc, which (sub)node is really populated: just the initial values, other subnodes ... ? Here I find that a verbose log file really could help ! -- Good luck Ivar
Muhammad Ikhlaq
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Dec 16, 2013, 6:19 a.m. EST
Hi,

i have a problem in transferring load from one physics to other i want to transfer load from Piezoelectric structure to FSI can you tell me which option i have to select for this purpose?

Best Regards


Hi

I have also issues to really understand when I get the initial conditions truely transferred, and when not,
I'm missing something like a line in the log file stating reading T from study 1 or something like that, and for each variable. I was woundering, if for the stress, if one shold add an initial stress sub node or not, I cannot find anything in the doc, which (sub)node is really populated: just the initial values, other subnodes ... ?

Here I find that a verbose log file really could help !

--
Good luck
Ivar


Hi, i have a problem in transferring load from one physics to other i want to transfer load from Piezoelectric structure to FSI can you tell me which option i have to select for this purpose? Best Regards [QUOTE] Hi I have also issues to really understand when I get the initial conditions truely transferred, and when not, I'm missing something like a line in the log file stating reading T from study 1 or something like that, and for each variable. I was woundering, if for the stress, if one shold add an initial stress sub node or not, I cannot find anything in the doc, which (sub)node is really populated: just the initial values, other subnodes ... ? Here I find that a verbose log file really could help ! -- Good luck Ivar [/QUOTE]
Sharareh Haddadi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 1, 2014, 4:29 a.m. EDT
hello
i have the same problem with acoustic-piezoelectric interaction with laminar flow and then using particle tracing as the results. but I can't couple acoustic with laminar flow and particle tracing. i can't see pressure and velocity field in the acoustophoresis force in particle tracing. I want to use pressure and velocity field that is made in acoustic-piezoelectric interaction physic in particle tracing but i can't couple them.
thanks
hello i have the same problem with acoustic-piezoelectric interaction with laminar flow and then using particle tracing as the results. but I can't couple acoustic with laminar flow and particle tracing. i can't see pressure and velocity field in the acoustophoresis force in particle tracing. I want to use pressure and velocity field that is made in acoustic-piezoelectric interaction physic in particle tracing but i can't couple them. thanks
Basky
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 10 years ago Jun 1, 2015, 10:13 a.m. EDT

Hi,

I have a problem to use a saved solution.

The model has two physics : Heat Transer and Solid Mechanics.

I disable the solver 1 and create an other solver. I select the solution of solver 1 as initial values for solver 2 by the dependent variables node : Method = solution
Solution = Solver 1
Time = automatic

It is working well for the temperature. But the stresses of Solver 2 at t=0 are 0!

Any help would be greatly appreciated :)

Alexis


Hi
i have the similar problem now. Did you figure out the solution for your problem?. It would be of great help to me please.....
[QUOTE] Hi, I have a problem to use a saved solution. The model has two physics : Heat Transer and Solid Mechanics. I disable the solver 1 and create an other solver. I select the solution of solver 1 as initial values for solver 2 by the dependent variables node : Method = solution Solution = Solver 1 Time = automatic It is working well for the temperature. But the stresses of Solver 2 at t=0 are 0! Any help would be greatly appreciated :) Alexis [/QUOTE] Hi i have the similar problem now. Did you figure out the solution for your problem?. It would be of great help to me please.....
Jason Lerch
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Jun 28, 2015, 9:57 a.m. EDT
I am having the same problem. I am using Pressure acoustics, transient and solid mechanics in a coupled model to see how the wave moves a water particle. Even though they are coupled, the solid mechanics model just shows 0 stress. The acoustics study runs fine and shows the expected result, but when I run a second study for the solid mechanics nothing happens.
I am having the same problem. I am using Pressure acoustics, transient and solid mechanics in a coupled model to see how the wave moves a water particle. Even though they are coupled, the solid mechanics model just shows 0 stress. The acoustics study runs fine and shows the expected result, but when I run a second study for the solid mechanics nothing happens.
Affar Karimullah
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Nov 25, 2015, 1:03 p.m. EST
I too am having problems with the following model. I have an electromagnetic heat source for which I run an EM waves physics, then a heat transfer in solids for looking at the heating effect in a plasmonic structure over 8ns. Now all of this works well however I need to measure the heat diffusion after the laser pulse, that is the EM source is turned off. Hence I created another heat transfer physics (ht2) component and set up study 2 to solve only for the ht2 only. My current problem is that I cannot transfer the temperature (T1) from ht1 to the temperature (T2) of ht2. While the values of T1 from study 1 do transfer into study 2 , I need to figure out how I can set them as the initial values for ht2's T2.
I too am having problems with the following model. I have an electromagnetic heat source for which I run an EM waves physics, then a heat transfer in solids for looking at the heating effect in a plasmonic structure over 8ns. Now all of this works well however I need to measure the heat diffusion after the laser pulse, that is the EM source is turned off. Hence I created another heat transfer physics (ht2) component and set up study 2 to solve only for the ht2 only. My current problem is that I cannot transfer the temperature (T1) from ht1 to the temperature (T2) of ht2. While the values of T1 from study 1 do transfer into study 2 , I need to figure out how I can set them as the initial values for ht2's T2.
Abdulrahman Al Rumaidh
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Dec 25, 2015, 4:58 a.m. EST

Dear Osameh,

The way to do this is to solve your initial model, then go back to Study 1. Open it, right-click "Solver 1" and press "Disable". This means that the solver sequence, and the solution attached to it, will not be changed. If you now right-click Study 1, and press "Show default solver", then a new solver sequence is built (Solver 2). Open this one, and in "Dependent variables", choose the solution from Solver 1 as initial values.

Tip #1: To keep an overview, it helps to rename "Solver 1" to "Solver V=3" or something related to your boundary condition, so that you can keep track of which solver contains what. The same goes for the solution.

Tip #2: You can automatize this process by using a parametric sweep.

Hope that helps,

Ruud Borger
COMSOLBV


Are these steps supposed to accelerate convergence? Because I followed them but I get a solution in the exact same time of the previous one.


--
Abdulrahman Al Rumaidh
Mechanical Engineering Student | University of Bahrain
[QUOTE] Dear Osameh, The way to do this is to solve your initial model, then go back to Study 1. Open it, right-click "Solver 1" and press "Disable". This means that the solver sequence, and the solution attached to it, will not be changed. If you now right-click Study 1, and press "Show default solver", then a new solver sequence is built (Solver 2). Open this one, and in "Dependent variables", choose the solution from Solver 1 as initial values. Tip #1: To keep an overview, it helps to rename "Solver 1" to "Solver V=3" or something related to your boundary condition, so that you can keep track of which solver contains what. The same goes for the solution. Tip #2: You can automatize this process by using a parametric sweep. Hope that helps, Ruud Borger COMSOLBV [/QUOTE] Are these steps supposed to accelerate convergence? Because I followed them but I get a solution in the exact same time of the previous one. -- Abdulrahman Al Rumaidh Mechanical Engineering Student | University of Bahrain

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