[Trilinos-Users] NOx-LOCA issue with convergence - Arc length continuation

[narendiran kumar]

hi,

Initially, I have solved no of equation for steady state solution with
constant parameter, and then i used that one as my initial guess (file:
steadystatevalue.txt) for solving arc length continuation.  Computation
should start from that initial guess (y1, y2, .... yn, p) but It's giving
the variable solution (y1, y2, ... ,yn) closure to zero with that actual
initial guess parameter value(p) in arc length continuation method.   What i
have calculated manual steady state solution is not same with this arc
length continuation steady state solution for particular parameter value.  I
have added below code for fixing intial guess (initial steady state value).
Is InitialGuess not properly initialized.


void
ChanProblemInterface::init() {

FILE *fp;
fp = fopen("steadystatevalue.txt", "rb");
int i;
float initialGuess[200];
i=0;
while(fscanf(fp, "%f", &initialGuess[i]) != EOF)
        {
         printf("%0.18f\n", initialGuess[i]);
         i=i+1;
       }
fclose(fp);
}



Regards,
K. Narendiran

On Tue, Jun 21, 2011 at 7:36 PM, Nico Schlömer <nico.schloemer at ua.ac.be>wrote:

> Hi Kumar,
>
> it'd be useful to know the parameters you use to drive the computations.
> Also, the tail'ed output you attached suggests that a number of
> continuation steps have been completed successfully; is that correct? The
> output there would also help debugging.
>
> Cheers,
> Nico
>
>
>
>
> On 06/21/2011 03:14 PM, narendiran kumar wrote:
>
>> hi,
>>
>> I am using nox package for solving vector problem (nof. eqn 200) using
>> arc-length continuation. I have modified chancontinuation example
>> program and i gave initial guess in vector form (for n = 200 eqn). When
>> i was compile, i am getting warring. I guess it's because of initial
>> guess. Can anyone please help me.
>>
>> thanks in advance,
>>
>> Regards,
>> K. Narendiran
>>
>>
>> Example warring o/p:
>>
>> **************************************************
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**
>> ~~~~~~~~~~~~
>> Start of Continuation Step 11 : Parameter: Da = 8.747e-01 from 9.511e-01
>> Continuation Method: Arc Length
>> Current step size = 8.434e-05 Previous step size = 5.812e-05
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**
>> ~~~~~~~~~~~~
>>
>> Warning: NOX::Solver::LineSearchBased::**init() - The solution passed
>> into
>> the solver (either through constructor or reset method) is already
>> converged! The solver wil not attempt to solve this system since status
>> is flagged as converged.
>>
>> ****************************************************************
>> ************
>> -- Nonlinear Solver Step 0 --
>> ||F|| = 8.598e+00 step = 0.000e+00 dx = 0.000e+00 (Converged!)
>> ****************************************************************
>> ************
>>
>> ****************************************************************
>> ************
>> -- Final Status Test Results --
>> Converged....OR Combination ->
>> Converged....F-Norm = 6.065e-01 < 1.000e+00
>> (Length-Scaled Two-Norm, Absolute Tolerance)
>> ??...........Number of Iterations = -1 < 20
>> ****************************************************************
>> ************
>>
>> ****************************************************************
>> ************
>> -- Nonlinear Solver Step 0 --
>> ||F|| = 8.598e+00 step = 0.000e+00 dx = 0.000e+00 (Converged!)
>> ****************************************************************
>> ************
>>
>> ****************************************************************
>> ************
>> -- Final Status Test Results --
>> Converged....OR Combination ->
>> Converged....F-Norm = 6.065e-01 < 1.000e+00
>> (Length-Scaled Two-Norm, Absolute Tolerance)
>> ??...........Number of Iterations = -1 < 20
>> ****************************************************************
>> ************
>>
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**
>> ~~~~~~~~~~~~
>> End of Continuation Step 11 : Parameter: Da = 8.747e-01 from 9.511e-01
>> --> Step Converged in 0 Nonlinear Solver Iterations!
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**
>> ~~~~~~~~~~~~
>>
>>
>> Calling Predictor with method: Secant
>>
>> ++++++++++++++++++++++++++++++**++++++++++++++++++++++++++++++**++++
>> Arc-length scaling calculation for parameter Da:
>> Parameter component of predictor before rescaling = 9.058e+02
>> Scale factor from previous step 1.000e-03
>> Parameter contribution to arc-length equation = 9.058e-01
>>
>> Parameter component of predictor after rescaling = 9.058e+02
>> New scale factor (theta) = 1.000e-03
>> Parameter contribution to arc-length equation = 9.058e-01
>> ++++++++++++++++++++++++++++++**++++++++++++++++++++++++++++++**++++
>> At parameter value: 0.87472249 the solution vector (norm=0.0096190908) is
>> 0.00045035754 0.0011253381 0.00036360756 0.00090869718 0.000277962
>> 0.00069468095 ... 0.0053114264 0.00023180336
>>
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~**
>> ~~~~~~~~~~~~
>>
>>
>> ********************************************
>>
>>
>>
>> ______________________________**_________________
>> Trilinos-Users mailing list
>> Trilinos-Users at software.**sandia.gov <Trilinos-Users at software.sandia.gov>
>> http://software.sandia.gov/**mailman/listinfo/trilinos-**users<http://software.sandia.gov/mailman/listinfo/trilinos-users>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://software.sandia.gov/pipermail/trilinos-users/attachments/20110621/f0096690/attachment.html