lenght, time and energy base scale in Crow app

[Giovar6]

Dear MOOSE users,

I'm here again with a problem with the Crow app, developed by @SudiptaBiswas. This time is more a conceptual problem.
In the material section of the input file, when calculating the mobility properties and the simulation coefficients, in the Crow application are specified some coefficients to non-dimensionalize the material properties.
To explain better what I'm referring to, I report an example of an input file developed for the Crow app. (I quote only the modules of the input file important to clarify what I am referring to)

[Mesh]
type = GeneratedMesh
dim = 2
nx = 80
ny = 40
nz = 0
xmin = 0.0
xmax = 400.0
ymin = 0.0
ymax = 200.0
zmax = 0
uniform_refine = 1
elem_type = QUAD4
[]

[Materials]
[./free_energy]
type = SinteringFreeEnergy
block = 0
c = c
v = 'gr0 gr1'
#A = A
#B = B
#f_name = S
derivative_order = 2
#outputs = console
[../]
[./coeff]
type = SinteringCoefficients
block = 0
T = 800.0
int_width = 20
GBmob0 = 2.5e-6
Qgbm = 0.23
surface_energy = 1.03e19
GB_energy = 4.42e18
length_scale = 1e-08
time_scale = 0.1
energy_scale = 100.0
energy_unit = eV
outputs = exodus
[../]
[./mob]
type = SinteringMtrxMobility
T = 800.0
int_width = 20
Qv = 2.19
Qvc = 2.19
Qgb = 0.75
Qs = 0.9
Dgb0 = 1.95e-8
Dsurf0 = 2.6e-5
Dvap0 = 7.8e-5
Dvol0 = 7.8e-5
c = c
v = 'gr0 gr1'
Vm = 1.182e-29
length_scale = 1e-08
time_scale = 0.1
bulkindex = 1.0
surfindex = 1.0
gbindex = 1.0
#prefactor = 0.1
outputs = exodus
[../]
[]

[Executioner]
Preconditioned JFNK (default)
type = Transient
scheme = BDF2
solve_type = NEWTON
petsc_options_iname = '-pc_type -ksp_grmres_restart -sub_ksp_type -sub_pc_type -pc_asm_overlap -sub_pc_factor_shift_type'
petsc_options_value = 'asm 31 preonly lu 1 nonzero'
#petsc_options = '-snes_converged_reason'
l_max_its = 20
nl_max_its = 20
nl_abs_tol = 1e-08
nl_rel_tol = 1e-06
l_tol = 1e-04
end_time = 5000
dt = 0.01
[./TimeStepper]
type = IterationAdaptiveDT
dt = 0.01
growth_factor = 1.2
[../]
[]

[ICs]
[./ic_gr1]
int_width = 20.0
x1 = 250.0
y1 = 100.0
radius = 74.0
outvalue = 0.0
variable = gr1
invalue = 1.0
type = SmoothCircleIC
[../]
[./multip]
x_positions = '100.0 250.0'
int_width = 20.0
z_positions = '0 0'
y_positions = '100.0 100.0 '
radii = '74.0 74.0'
3D_spheres = false
outvalue = 0.001
variable = c
invalue = 0.999
type = SpecifiedSmoothCircleIC
block = 0
[../]
[./ic_gr0]
int_width = 20.0
x1 = 100.0
y1 = 100.0
radius = 74.0
outvalue = 0.0
variable = gr0
invalue = 1.0
type = SmoothCircleIC
[../]
[]

In this case are are adopted as length scale 1e-08, time scale 0.1 and energy scale 100. The problem is that i cannot understand how this exact values are identified. Why I cannot use for example 1e-06?
I have tried to perform simulations with these values, but the simulation did not converged. So I understand that is important to specify the correct values, but how?

Thank you in advance to anyone will give me any suggestion or advice.
Giovanni Rizza

[aeslaughter]

@SudiptaBiswas This is super old, but curious if you had any advice.

[SudiptaBiswas]

@Giovar6 The scales are chosen such that the nondimensional mobility values are close to unity. This is done for computational benefits. This is generic for phase-field models, nothing specific to Crow. Feel free to use different scales that make sense for your simulations.

[milljm]

Is there a way you can post text as text instead of images of text? I can't for example, easily clone the version of PETSc you are grabbing... because I can't select text from a picture.

What version of MOOSE did you end up cloning? Perhaps that version of MOOSE is too old to make use of newer compilers. I'd like to try for myself...

[HPL001]

I cloned this version of moose (https://github.com/yardasol/moose.git), and this version of Crow (https://github.com/luchinsky/Crow.git). Some of the information in the figure is in the txt file.
001.txt

[milljm]

I need you to enter the MOOSE directory and run:

git rev-parse HEAD

This will tell me what hash you checked out.

[HPL001]

ok, this is the output.
002.txt

[HPL001]

@milljm Hello, dear developer. Are there any solutions now? Looking forward to your reply.