What does the output mean?

[domcivocato]

Hello fellow Simulators,

i wanted to ask you what the numbers in my output mean. I am a little confused in decrypting what it means.
Mostly what i dont understand is the variable "nrm". Also when you see in my log file the numbers that kind of repeat themselves.
For example 10 0.7514E+00 etc.

If you have a clue, let me know.

Thank you,
domcivocato

StatCurrentSolver: ------------------------------------------------
StatCurrentSolver: Solving static current conduction solver
StatCurrentSolverBulkAssembly: Elapsed REAL time: 0.0779 (s)
StatCurrentSolverBCAssembly: Elapsed REAL time: 0.0026 (s)
CRS_IncompleteLU: ILU(0) (Real), Performing Factorization:
CRS_IncompleteLU: ILU(0) (Real), NOF nonzeros: 49473
CRS_IncompleteLU: ILU(0) (Real), filling (%) : 100
CRS_IncompleteLU: ILU(0) (Real), Factorization ready at (s): 0.00
10 0.7514E+00
20 0.4667E+00
30 0.2966E+00
40 0.1546E+00
50 0.1174E+00
60 0.1010E+00
70 0.2172E+00
80 0.1123E+00
90 0.1352E+00
100 0.2056E+00
110 0.2311E+00
120 0.3243E+00
130 0.4547E+00
140 0.6345E+00
150 0.6753E+00
160 0.9078E+00
170 0.1077E+01
180 0.1010E+01
190 0.8410E+00
200 0.8324E+00
210 0.8051E+00
220 0.3766E+00
230 0.3817E+00
240 0.3720E+00
250 0.3660E+00
260 0.4157E+00
270 0.4478E+00
280 0.1279E+02
290 0.1471E+01
300 0.2861E+01
310 0.2783E+01
320 0.4185E+02
330 0.5257E+01
340 0.2888E+01
350 0.2737E+01
360 0.2365E+01
370 0.2394E+01
380 0.2201E+01
390 0.2503E+01
400 0.5770E+01
410 0.5320E+01
420 0.5224E+01
430 0.5260E+01
440 0.5682E+01
450 0.5841E+01
460 0.4841E+01
470 0.4971E+01
480 0.3135E+01
490 0.3263E+01
500 0.3249E+01
501 0.3249E+01
IterSolve: Linear iteration did not converge to tolerance
ComputeChange: NS (ITER=1) (NRM,RELC): ( 4874034.9 2.0000000 ) :: currentsolver

[Rich_B]

Hello,

Looking at the end of the solver log, where the heat solver doesn't converge, these rows appear:

IterSolve: Linear iteration did not converge to tolerance
ComputeChange: NS (ITER=2) (NRM,RELC): ( 0.38565092E+37 1.9979994 ) :: heat equation
HeatSolve: iter: 2 Assembly: (s) 0.14 0.28
HeatSolve: iter: 2 Solve: (s) 0.27 0.53
HeatSolve: Result Norm : 3.8565092019157088E+036
HeatSolve: Relative Change : 1.9979994462034139

So NRM = Result Norm, and RELC = Relative Change

The solution doesn't converge. If you want help with convergence, post the geometry file, instructions on meshing, and the sif file.

Rich.

[domcivocato]

Hello Rich,

thanks for your answer! Now i get the abbreviations.
Can you make sense of the values posted before? Like

480 0.3135E+01
490 0.3263E+01
500 0.3249E+01
501 0.3249E+01

This seems also like a time or iteration step measure for the convergence. But this deviates from the convergence that is depicted in the RELC.
Also the time or iteration step seems weird, because i didnt specify nothing in the range of 500. Do you maybe have a clue what this numbers stand for?

Cheers,
Dominik

[kevinarden]

The default iterative solver generated by the gui usually has something like this
Nonlinear System Convergence Tolerance = 1.0e-7
Nonlinear System Max Iterations = 20
Nonlinear System Newton After Iterations = 3
Nonlinear System Newton After Tolerance = 1.0e-3
Nonlinear System Relaxation Factor = 1
Linear System Solver = Iterative
Linear System Iterative Method = BiCGStab
Linear System Max Iterations = 500
Linear System Convergence Tolerance = 1.0e-10

the output
480 0.3135E+01
490 0.3263E+01
500 0.3249E+01
501 0.3249E+01

is from the linear solver and it goes until it hits the convergence tolerance
or the max iterations

The other output is from the non-linear solver, the code alternates between the linear solver and the non-linear solver until it converges on the steady state tolerance or it diverges.

[domcivocato]

Thanks for your answer, that makes sense!

Cheers,
Dominik

[raback]

Hi Dominik,

The 1st thing to get convergence here could be to change "ILU0" to "ILU1" or "ILU2". From linear solvers "BiCGtabl" and "idrs" are usually the most robust.

-Peter

[domcivocato]

Okay thanks, that is very good to know.

Cheers,
Dominik