Dear Ivar,
We are solving a nonlinear transient problem using a "fully coupled approach" with the constant Newton option (damping factor =1).
The termination criterion is the default tolerance with a maximum of 6 Newton iterations.
The direct solver used is the default MUMPS.
We set the detailed log on, to see the steps the solver takes.
We are trying to understand how the time step stepping and non-linear solving are proceeding.
For a given time step (say 17), we did not reach convergence after the 6 Newton iterations (iter 1 to 6).
(See the log copy below)
Then, The non-linear solver takes a NEW step (iter 1) and immediately converges, such that the general time solver can take a new time step (18).
What happened ?
What is the procedure when the newton iterative solver did not reach convergence after 6 iterations ?
Is the time step size reduced to accommodate for the detected non-linearity ?
We could not find any information in the documentation about what happens if a non-linear solver reaches the maximum number of iterations.
thank you for your help,
regards
Arthur
-----------------------------
LOG WINDOW:
Step Time Stepsize Res Jac Sol Order Tfail NLfail
17 0.0016168 0.00012558 61 24 60 1 0 2
Time t = 0.00161681.
Iter Damping Stepsize #Res #Jac #Sol
1 1.0000000 0.0035 62 25 61
2 1.0000000 1.5e+002 63 25 62
3 1.0000000 3.9e+009 64 25 63
4 1.0000000 8.1e+020 65 25 64
5 1.0000000 1.9e+062 66 25 65
6 1.0000000 1.$ 67 25 66
Iter Damping Stepsize #Res #Jac #Sol
1 1.0000000 0.00029 68 26 67
Step Time Stepsize Res Jac Sol Order Tfail NLfail
18 0.0016796 6.2789e-005 68 26 67 1 0 3
---------------------------------------
We are solving a nonlinear transient problem using a "fully coupled approach" with the constant Newton option (damping factor =1).
The termination criterion is the default tolerance with a maximum of 6 Newton iterations.
The direct solver used is the default MUMPS.
We set the detailed log on, to see the steps the solver takes.
We are trying to understand how the time step stepping and non-linear solving are proceeding.
For a given time step (say 17), we did not reach convergence after the 6 Newton iterations (iter 1 to 6).
(See the log copy below)
Then, The non-linear solver takes a NEW step (iter 1) and immediately converges, such that the general time solver can take a new time step (18).
What happened ?
What is the procedure when the newton iterative solver did not reach convergence after 6 iterations ?
Is the time step size reduced to accommodate for the detected non-linearity ?
We could not find any information in the documentation about what happens if a non-linear solver reaches the maximum number of iterations.
thank you for your help,
regards
Arthur
-----------------------------
LOG WINDOW:
Step Time Stepsize Res Jac Sol Order Tfail NLfail
17 0.0016168 0.00012558 61 24 60 1 0 2
Time t = 0.00161681.
Iter Damping Stepsize #Res #Jac #Sol
1 1.0000000 0.0035 62 25 61
2 1.0000000 1.5e+002 63 25 62
3 1.0000000 3.9e+009 64 25 63
4 1.0000000 8.1e+020 65 25 64
5 1.0000000 1.9e+062 66 25 65
6 1.0000000 1.$ 67 25 66
Iter Damping Stepsize #Res #Jac #Sol
1 1.0000000 0.00029 68 26 67
Step Time Stepsize Res Jac Sol Order Tfail NLfail
18 0.0016796 6.2789e-005 68 26 67 1 0 3
---------------------------------------