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 The information provided may be out of date.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.

Questions about the Log while solving

Please login with a confirmed email address before reporting spam

Hi again,

I am now running a model that I have built. It has been running for 3 days now(don't worry, I expected this - each calculation takes around 8 hours~ and I gave it a parametric sweep so it has a lot to do), and I want to try and understand how the calculation progresses while it is running.

the parametric sweep parameter was the current. so at first it started with the first value and moved on to the next etc. on one occasion, it seems to have failed to calculate it, so it came between - i.e. it solved 3000A and moved on to 4000, then returned to 3500 and to 3250. after it succeeded in 3250 it went on to 3411 and 3733. anyway, my question is about the log file - I want to understand what happened/is happening; can anyone help me here?

The following is copied-pasted from the log:


Parameter J0 = 3000.
Iter ErrEst Damping Stepsize #Res #Jac #Sol LinIt LinErr LinRes
1 2.7e+02 0.0010000 2.7e+02 5 1 6 57 0.038 9.5e-05
2 1.4e+02 0.0010000 1.4e+02 7 2 9 83 0.026 6.5e-05
3 65 0.0036038 65 8 3 11 111 0.082 0.0002
4 7.9 0.0222547 8.1 9 4 13 139 0.069 0.00017
5 1.8 0.2225474 2.2 10 5 15 159 0.1 0.00026
6 0.24 0.6856506 0.78 11 6 17 179 0.11 0.00026
7 0.00052 1.0000000 0.2 12 7 20 202 0.045 0.00011
8 0.00034 1.0000000 0.0016 13 8 22 210 0.12 0.0003

Parameter J0 = 4000.
Iter ErrEst Damping Stepsize #Res #Jac #Sol LinIt LinErr LinRes

Parameter J0 = 3500.
Iter ErrEst Damping Stepsize #Res #Jac #Sol LinIt LinErr LinRes

Parameter J0 = 3250.
Iter ErrEst Damping Stepsize #Res #Jac #Sol LinIt LinErr LinRes
1 0.018 0.1000000 0.02 20 11 31 28114 0.12 0.0003
2 0.017 0.1000000 0.019 22 12 35 28212 0.081 0.0002
3 0.016 0.1331932 0.018 24 13 39 28280 0.12 0.00031
4 0.014 0.1924893 0.018 26 14 43 28344 0.11 0.00029
5 0.014 1.0000000 0.015 27 15 46 28398 0.11 0.00027
6 0.0032 1.0000000 0.0038 28 16 48 28416 0.11 0.00027
7 0.0016 0.7944232 0.0037 29 17 50 28438 0.12 0.00031
8 0.00092 0.7271127 0.0032 30 18 53 28460 0.11 0.00026
9 6.3e-05 1.0000000 0.0018 31 19 55 28480 0.12 0.0003

Parameter J0 = 3411.05.
Iter ErrEst Damping Stepsize #Res #Jac #Sol LinIt LinErr LinRes
1 0.00063 1.0000000 0.0017 34 20 59 28518 0.12 0.00031


right now it's running on J0=3733.16 at 8%, with the FMGRES at 74%, convergence of 1.06, if that's relevant..



So I wanted to know what these numbers mean? what's the Damping, Stepsize, #res etc?..


Sorry for this post's clumsiness.. but thanks anyway :)

8 Replies Last Post Jan 15, 2012, 11:02 p.m. EST
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Nov 30, 2011, 10:40 a.m. EST
Hi

I do not know all the parameters from the solver, but what you see is that the solver made a large step, then was not happy with the convergence, and reduced the step, tried again, failed, then reduced it again. This is typical for iterative solvers, I'm not sure how much of the results of the larger steps are used for defining next smaller step, but the last valid steps are at least used to project to the next solver step.

This is the typical behaviour of the iterative solver, in "free" mode.

You can also force it to do only your defined steps, but then there is a fair chance that it will just stop when it discovers that it is diverging. check the lower level solver sub-nodes

--
Good luck
Ivar
Hi I do not know all the parameters from the solver, but what you see is that the solver made a large step, then was not happy with the convergence, and reduced the step, tried again, failed, then reduced it again. This is typical for iterative solvers, I'm not sure how much of the results of the larger steps are used for defining next smaller step, but the last valid steps are at least used to project to the next solver step. This is the typical behaviour of the iterative solver, in "free" mode. You can also force it to do only your defined steps, but then there is a fair chance that it will just stop when it discovers that it is diverging. check the lower level solver sub-nodes -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 5, 2012, 9:59 a.m. EST
Hi again,

I am once again trying to understand the information that is given to me in the Log tab.

my log is here, every space was replaced with _ so that it will be readable. (sorry for the long wall of text)


Step________Time____Stepsize______Res__Jac__Sol_Order_Tfail_NLfail_LinIt___LinErr___LinRes
___0___________0_____________out____4____6____4__________________0
___________________Group_#1:________2____3____2_______________________14___0.0091__0.00022
___________________Group_#2:________2____3____2________________________4__0.00037__1.9e-06
___1____0.031416____0.031416________6____8____6_____1_____0______0
___________________Group_#1:________3____4____3_______________________21___0.0086__0.00056
___________________Group_#2:________3____4____3________________________6___0.0039__3.3e-06
___2____0.062832____0.031416________8____8____8_____1_____0______0
___________________Group_#1:________4____4____4_______________________32___0.0098__0.00045
___________________Group_#2:________4____4____4________________________9___0.0011__7.2e-07
___3_____0.12566____0.062832_______10____8___10_____2_____0______0
___________________Group_#1:________5____4____5_______________________42___0.0073__0.00037
___________________Group_#2:________5____4____5_______________________11___0.0087__8.8e-06
___4_____0.25133_____0.12566_______12___10___12_____2_____0______0
___________________Group_#1:________6____5____6_______________________51___0.0063__0.00019
___________________Group_#2:________6____5____6_______________________14___0.0015__1.3e-06
___5_____0.50265_____0.25133_______16___12___16_____2_____0______0
___________________Group_#1:________8____6____8_______________________58___0.0056__1.1e-06
___________________Group_#2:________8____6____8_______________________18__7.8e-09__1.8e-11

...


from what I understand, the step is clearly the step number; time is the current value of the t variable; stepsize is the stepsize of the time jumps in every step. these are the easy ones of course. my question is about the next ones, if anyone can help me.

in the step size column there are always the group #1 and 2, which to my understanding, gives me the information for the each group in the segregated solver.

about the Res, Jac, Sol, I don't understand it at all; the Sol and Res columns are exactly the same, so I suppose that the res is some result of a computation and that the sol is the final value, which happens to be the same. I don't know what the Jac means(I'm guessing it's jacobian somewhere but I don't know where/how it is relevant).

then there are the Order, Tfail and NLfail. what do they mean?
likewise about the LinIt, LinRes, and LinErr. I guess that the 'It' refers to iteration, Res is result and Err is clearly error - but what is that Lin which is talked about?

I want to be able to understand the log and conclude about how the computation is progressing - please help :(

it should be noted that the solver is the default for Induction Heating in a Time Dependent model.


Thank you very much for your time,
Shoval
Hi again, I am once again trying to understand the information that is given to me in the Log tab. my log is here, every space was replaced with _ so that it will be readable. (sorry for the long wall of text) Step________Time____Stepsize______Res__Jac__Sol_Order_Tfail_NLfail_LinIt___LinErr___LinRes ___0___________0_____________out____4____6____4__________________0 ___________________Group_#1:________2____3____2_______________________14___0.0091__0.00022 ___________________Group_#2:________2____3____2________________________4__0.00037__1.9e-06 ___1____0.031416____0.031416________6____8____6_____1_____0______0 ___________________Group_#1:________3____4____3_______________________21___0.0086__0.00056 ___________________Group_#2:________3____4____3________________________6___0.0039__3.3e-06 ___2____0.062832____0.031416________8____8____8_____1_____0______0 ___________________Group_#1:________4____4____4_______________________32___0.0098__0.00045 ___________________Group_#2:________4____4____4________________________9___0.0011__7.2e-07 ___3_____0.12566____0.062832_______10____8___10_____2_____0______0 ___________________Group_#1:________5____4____5_______________________42___0.0073__0.00037 ___________________Group_#2:________5____4____5_______________________11___0.0087__8.8e-06 ___4_____0.25133_____0.12566_______12___10___12_____2_____0______0 ___________________Group_#1:________6____5____6_______________________51___0.0063__0.00019 ___________________Group_#2:________6____5____6_______________________14___0.0015__1.3e-06 ___5_____0.50265_____0.25133_______16___12___16_____2_____0______0 ___________________Group_#1:________8____6____8_______________________58___0.0056__1.1e-06 ___________________Group_#2:________8____6____8_______________________18__7.8e-09__1.8e-11 ... from what I understand, the step is clearly the step number; time is the current value of the t variable; stepsize is the stepsize of the time jumps in every step. these are the easy ones of course. my question is about the next ones, if anyone can help me. in the step size column there are always the group #1 and 2, which to my understanding, gives me the information for the each group in the segregated solver. about the Res, Jac, Sol, I don't understand it at all; the Sol and Res columns are exactly the same, so I suppose that the res is some result of a computation and that the sol is the final value, which happens to be the same. I don't know what the Jac means(I'm guessing it's jacobian somewhere but I don't know where/how it is relevant). then there are the Order, Tfail and NLfail. what do they mean? likewise about the LinIt, LinRes, and LinErr. I guess that the 'It' refers to iteration, Res is result and Err is clearly error - but what is that Lin which is talked about? I want to be able to understand the log and conclude about how the computation is progressing - please help :( it should be noted that the solver is the default for Induction Heating in a Time Dependent model. Thank you very much for your time, Shoval

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 5, 2012, 10:41 a.m. EST
Hi

welcome to the club ;)

Well I would also be interested to better understand some of those.

one you missed is the "out" that indicates that COMSOL has interpolated for a user requested time (or parameter) step and saved the related solution data set part

By the way there is a way to get the solver to tell even more (not much but info on scaling etc) check the solver sub nodes "logfile: detailed

--
Good luck
Ivar
Hi welcome to the club ;) Well I would also be interested to better understand some of those. one you missed is the "out" that indicates that COMSOL has interpolated for a user requested time (or parameter) step and saved the related solution data set part By the way there is a way to get the solver to tell even more (not much but info on scaling etc) check the solver sub nodes "logfile: detailed -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 5, 2012, 11:22 a.m. EST
ahh yes, the out is something that I figured as well, it appears more in the next parts of the log and exactly at time=pi/4 *n, which is the times I asked for the interpolation, so I understood this :)

is there some dynamic help about this, or anything in the tutorial?
ahh yes, the out is something that I figured as well, it appears more in the next parts of the log and exactly at time=pi/4 *n, which is the times I asked for the interpolation, so I understood this :) is there some dynamic help about this, or anything in the tutorial?

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 5, 2012, 3:24 p.m. EST
The log file variables are explained in the documentation. See COMSOL Multiphysics > Solvers and Study Types > Solver Studies and Study Types. I only found this documentation section recently, so either they added it recently or I was just looking in the wrong place.

Nagi Elabbasi
Veryst Engineering
The log file variables are explained in the documentation. See COMSOL Multiphysics > Solvers and Study Types > Solver Studies and Study Types. I only found this documentation section recently, so either they added it recently or I was just looking in the wrong place. Nagi Elabbasi Veryst Engineering

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 6, 2012, 1:17 a.m. EST
Hi, Happy New Year,
and thanks for the link Nagi

I found more (or the same) in the help under search "Progress and Log Information"

--
Good luck
Ivar
Hi, Happy New Year, and thanks for the link Nagi I found more (or the same) in the help under search "Progress and Log Information" -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 15, 2012, 8:14 a.m. EST
Hi again,

firstly I would like to thank you for the information you provided above, it proved helpful :)

second, I have a new question related to this subject. Once again I will post a partial quote of the log:

Warning: Inverted mesh element near coordinate(0.4,-0.478926,-1.57896).

This line appears around 50 times in my log file now; it appears almost after every line, multiple times. Why does it report so much about the same thing? do you know what it means?

The computation is running for almost 80 hours now, so I am not stopping it; but it seems wierd - especially since my model is very small, and the Z coordinate -1.57 is far away from the place where the physics matters. what I'm trying to say is that my model is around 1mX70cm on area and around 12 cm on height(Z coordinate), and around it there's a sphere of air in radius of 2m, so what's wrong with that specific coordinate?


Thank you very much once again for your answers and time,
Shoval
Hi again, firstly I would like to thank you for the information you provided above, it proved helpful :) second, I have a new question related to this subject. Once again I will post a partial quote of the log: Warning: Inverted mesh element near coordinate(0.4,-0.478926,-1.57896). This line appears around 50 times in my log file now; it appears almost after every line, multiple times. Why does it report so much about the same thing? do you know what it means? The computation is running for almost 80 hours now, so I am not stopping it; but it seems wierd - especially since my model is very small, and the Z coordinate -1.57 is far away from the place where the physics matters. what I'm trying to say is that my model is around 1mX70cm on area and around 12 cm on height(Z coordinate), and around it there's a sphere of air in radius of 2m, so what's wrong with that specific coordinate? Thank you very much once again for your answers and time, Shoval

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 15, 2012, 11:02 p.m. EST
You have a bad element at the specified location which would put it in the air region. The solver does not keep track, it seems, of the mesh warnings that it reported in previous calls so it will give you this message more than you care to see it.

Nagi Elabbasi
Veryst Engineering
You have a bad element at the specified location which would put it in the air region. The solver does not keep track, it seems, of the mesh warnings that it reported in previous calls so it will give you this message more than you care to see it. Nagi Elabbasi Veryst Engineering

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.