Issues with convergence in simulations containing 2 phase

[deviantan]

Hi, I was working on a simulation of a dual phase material which resulted in non-convergence of the defined criteria. Then I tried solving the worked examples -- (C:\FEPX\tests\05_dual_bcc_fcc) with same .cfg and .msh file provided in the tests folder. Here also similar non-convergence was observed. Can you please let me know what I can do to overcome this issue.

Info   : A finite element software package for  polycrystal plasticity.
Info   : Version 2.1.0
Info   : Running on 1 cores.
Info   : <https://fepx.info>
Info   : Copyright (C) 1996-2023, DPLab, ACME Lab, CNRS.
Info   : ---------------------------------------------------------------
Info   : Start time: 2024-10-25 at 18:02
Info   : Loading simulation...
Info   :   [i] Parsing file `simulation.cfg'...
Info   :   - Material parameters:
Info   :     > Number of phases: 2
Info   :     > phase 1 - crystal type:   fcc
Info   :     > m:               0.200000E-01
Info   :     > gammadot_0:      0.100000E+01
Info   :     > h_0:             0.391900E+03
Info   :     > g_0:             0.200000E+03
Info   :     > g_s0:            0.335000E+03
Info   :     > n:               0.100000E+01
Info   :     > c11:             0.204600E+06
Info   :     > c12:             0.137700E+06
Info   :     > c44:             0.126200E+06
Info   :     > phase 2 - crystal type:   bcc
Info   :     > m:               0.200000E-01
Info   :     > gammadot_0:      0.100000E+01
Info   :     > h_0:             0.391900E+03
Info   :     > g_0:             0.200000E+03
Info   :     > g_s0:            0.335000E+03
Info   :     > n:               0.100000E+01
Info   :     > c11:             0.236900E+06
Info   :     > c12:             0.140600E+06
Info   :     > c44:             0.116000E+06
Info   :   [i] Parsed file `simulation.cfg'.
Info   :   [i] Parsing file `simulation.msh'...
Info   :   - Mesh parameters:
Info   :     > Node number: 49
Info   :     > Elt  number: 16
Info   :   [i] Parsed file `simulation.msh'.
Info   : Initializing simulation...
Info   :   - Initializing fields from isotropic viscoplastic solution
Info   :     > solveit_vp: Iteration 1
Info   :       . Solving NL iteration... R = 0.2639E-01 (31 iters)
Info   :     > solveit_vp: Iteration 2
Info   :       . Solving NL iteration... R = 0.3022E-02 (22 iters)
Info   :     > solveit_vp: Iteration 3
Info   :       . Solving NL iteration... R = 0.2576E-02 (22 iters)
Info   :     > Converged in 3 iterations
Info   : Running step 1...
Info   :   - Increment 1: t = 0.2500 secs, dt = 0.2500 secs
Info   :     > solveit_evp: Iteration 1
Warning:       . Error in line search
Warning:       . solve_stress_evp_newton: Some elements did not converge
Warning:       . solve_evp_stress: Some elements did not converge
Warning:       . Error in line search
Warning:       . solve_stress_evp_newton: Some elements did not converge
Warning:       . solve_evp_stress: Some elements did not converge
Warning:       . Error in line search

[deviantan]

Still not able to understand the issues with the simulation of dual phase models. Has anyone faced similar issue?

[mkasemer]

Apologies for the slow response. I am currently traveling, but I was unable to reproduce on my mac laptop. @rquey Do you have a similar error?

[rquey]

Yes. mpirun -np 2 fepx works fine, but mpirun -np 1 fepx doesn't:

==========================     F   E   P   X   ==========================
Info   : A finite element software package for  polycrystal plasticity.
Info   : Version 2.1.0                                             
Info   : Running on 1 cores.
Info   : <https://fepx.info>
Info   : Copyright (C) 1996-2023, DPLab, ACME Lab, CNRS.
Info   : ---------------------------------------------------------------
Info   : Start time: 2024-11-14 at 19:18
Info   : Loading simulation...
Info   :   [i] Parsing file `simulation.cfg'...
Info   :   - Material parameters:
Info   :     > Number of phases: 2
Info   :     > phase 1 - crystal type:   fcc 
Info   :     > m:               0.200000E-01
Info   :     > gammadot_0:      0.100000E+01
Info   :     > h_0:             0.391900E+03
Info   :     > g_0:             0.200000E+03
Info   :     > g_s0:            0.335000E+03
Info   :     > n:               0.100000E+01
Info   :     > c11:             0.204600E+06
Info   :     > c12:             0.137700E+06
Info   :     > c44:             0.126200E+06
Info   :     > phase 2 - crystal type:   bcc 
Info   :     > m:               0.200000E-01
Info   :     > gammadot_0:      0.100000E+01
Info   :     > h_0:             0.391900E+03
Info   :     > g_0:             0.200000E+03
Info   :     > g_s0:            0.335000E+03
Info   :     > n:               0.100000E+01
Info   :     > c11:             0.236900E+06
Info   :     > c12:             0.140600E+06
Info   :     > c44:             0.116000E+06
Info   :   [i] Parsed file `simulation.cfg'.
Info   :   [i] Parsing file `simulation.msh'...
Info   :   - Mesh parameters:
Info   :     > Node number: 49
Info   :     > Elt  number: 16
Info   :   [i] Parsed file `simulation.msh'.
Info   : Initializing simulation...
Info   :   - Initializing fields from isotropic viscoplastic solution
Info   :     > solveit_vp: Iteration 1
Info   :       . Solving NL iteration... R = 0.2639E-01 (31 iters)
Info   :     > solveit_vp: Iteration 2
Info   :       . Solving NL iteration... R = 0.3022E-02 (22 iters)
Info   :     > solveit_vp: Iteration 3
Info   :       . Solving NL iteration... R = 0.2576E-02 (22 iters)
Info   :     > Converged in 3 iterations
Info   : Running step 1...
Info   :   - Increment 1: t = 0.2500 secs, dt = 0.2500 secs
Info   :     > solveit_evp: Iteration 1
Warning:       . Error in line search
Warning:       . solve_stress_evp_newton: Some elements did not converge
Warning:       . solve_evp_stress: Some elements did not converge
Info   :       . Solving SA iteration... R = 0.2027E+00 (38 iters)
Info   :     > solveit_evp: Iteration 2
Warning:       . Error in line search
Warning:       . solve_stress_evp_newton: Some elements did not converge
Warning:       . solve_evp_stress: Some elements did not converge
Warning:       . Error in line search
Warning:       . solve_stress_evp_newton: Some elements did not converge
Warning:       . solve_evp_stress: Some elements did not converge
Warning:       . Error in line search
Warning:       . solve_stress_evp_newton: Some elements did not converge
Warning:       . solve_evp_stress: Some elements did not converge
Warning:       . Error in line search
Warning:       . solve_stress_evp_newton: Some elements did not converge
Warning:       . solve_evp_stress: Some elements did not converge
Warning:       . Error in line search
Warning:       . solve_stress_evp_newton: Some elements did not converge
Warning:       . solve_evp_stress: Some elements did not converge

[mkasemer]

Aha, good eye!

@deviantan: the short answer is to run in parallel. There are realistically no problems which will be run in serial at this point.

However, thank you for bringing this bug to our attention.