Error with the finer mesh from a Gmsh file #29113
Answered
by
GiudGiud
JuliaHLee
asked this question in
Q&A Meshing
-
|
Dear MOOSE experts, After MOOSE update, I keep getting an error when I run my models with gmsh file, which did run well without any problem before. While I was testing with a simple model with a squared domain including a circle in the middle, I found out that the model runs with coarse mesh at the boundary of a sideset block. But when I increase the number of nodes at the sideset boundary, it makes an error regarding the mesh. And the similar error appears when I run my models, which worked fine before. Could you help me with the issue? Please find the error, mesh file script and input file below. Many thanks Gmsh file script (circle.msh)Input file# Water properties
mu = 1.0E-3
rho = 1000.0
k = 0.598
cp = 4186
# Solid properties
cp_s = 830
rho_s = 1680
k_s = 3.5
# Other parameters
p_outlet = 0
u_inlet = -1e-4
h_conv = 50
[Mesh]
construct_side_list_from_node_list=true
[mesh_file]
type = FileMeshGenerator
file = circle.msh
[]
[sideset]
input = mesh_file
type = SideSetsBetweenSubdomainsGenerator
primary_block = domain
paired_block = 'c'
new_boundary = sideset
[]
[]
[GlobalParams]
rhie_chow_user_object = 'rc'
advected_interp_method = 'upwind'
velocity_interp_method = 'rc'
[]
[UserObjects]
[rc]
type = INSFVRhieChowInterpolator
u = vel_x
v = vel_y
pressure = pressure
block = domain
[]
[]
[Variables]
[vel_x]
type = INSFVVelocityVariable
initial_condition = 1e-4
block = domain
[]
[vel_y]
type = INSFVVelocityVariable
initial_condition = 1e-4
block = domain
[]
[pressure]
type = INSFVPressureVariable
block = domain
[]
[T]
type = INSFVEnergyVariable
initial_condition = 283.15
scaling = 1e-5
block = domain
[]
[Ts]
type = INSFVEnergyVariable
initial_condition = 333.15
scaling = 1e-5
block = 'c'
[]
[]
[FVKernels]
[mass]
type = INSFVMassAdvection
variable = pressure
rho = ${rho}
block = domain
[]
[u_time]
type = INSFVMomentumTimeDerivative
variable = vel_x
rho = ${rho}
momentum_component = 'x'
block = domain
[]
[u_advection]
type = INSFVMomentumAdvection
variable = vel_x
rho = ${rho}
momentum_component = 'x'
block = domain
[]
[u_viscosity]
type = INSFVMomentumDiffusion
variable = vel_x
mu = ${mu}
momentum_component = 'x'
block = domain
[]
[u_pressure]
type = INSFVMomentumPressure
variable = vel_x
momentum_component = 'x'
pressure = pressure
block = domain
[]
[v_time]
type = INSFVMomentumTimeDerivative
variable = vel_y
rho = ${rho}
momentum_component = 'y'
block = domain
[]
[v_advection]
type = INSFVMomentumAdvection
variable = vel_y
rho = ${rho}
momentum_component = 'y'
block = domain
[]
[v_viscosity]
type = INSFVMomentumDiffusion
variable = vel_y
mu = ${mu}
momentum_component = 'y'
[]
[v_pressure]
type = INSFVMomentumPressure
variable = vel_y
momentum_component = 'y'
pressure = pressure
block = domain
[]
[energy_time]
type = INSFVEnergyTimeDerivative
variable = T
rho = ${rho}
dh_dt = dh_dt
block = domain
[]
[temp_conduction]
type = FVDiffusion
coeff = 'k'
variable = T
block = domain
[]
[temp_advection]
type = INSFVEnergyAdvection
variable = T
block = domain
[]
[solid_energy_time]
type = INSFVEnergyTimeDerivative
variable = Ts
rho = ${rho_s}
dh_dt = dh_solid_dt
block = 'c'
[]
[solid_temp_conduction]
type = FVDiffusion
coeff = 'k_s'
variable = Ts
block = 'c'
[]
[]
[FVInterfaceKernels]
[convection]
type = FVConvectionCorrelationInterface
variable1 = T
variable2 = Ts
subdomain1 = domain
subdomain2 = 'c'
boundary = sideset
h = ${h_conv}
T_solid = Ts
T_fluid = T
wall_cell_is_bulk = true
[]
[]
[FVBCs]
[inlet-u]
type = INSFVInletVelocityBC
boundary = 'top'
variable = vel_x
function = 0
[]
[inlet-v]
type = INSFVInletVelocityBC
boundary = 'top'
variable = vel_y
function = ${u_inlet}
[]
[inlet_T]
type = FVDirichletBC
variable = T
boundary = 'top'
value = 283.15
[]
[no-slip-u]
type = INSFVNoSlipWallBC
boundary = 'left right sideset'
variable = vel_x
function = 0
[]
[no-slip-v]
type = INSFVNoSlipWallBC
boundary = 'left right sideset'
variable = vel_y
function = 0
[]
[outlet_p]
type = INSFVOutletPressureBC
boundary = 'bottom'
variable = pressure
function = '${p_outlet}'
[]
[]
[FunctorMaterials]
[functor_constants]
type = ADGenericFunctorMaterial
prop_names = 'cp k'
prop_values = '${cp} ${k}'
block = domain
[]
[ins_fv]
type = INSFVEnthalpyFunctorMaterial
temperature = 'T'
rho = ${rho}
block = domain
[]
[solid_functor_constants]
type = ADGenericFunctorMaterial
prop_names = 'cp_s k_s'
prop_values = '${cp_s} ${k_s}'
block = 'c'
[]
[solid_ins_fv]
type = INSFVEnthalpyFunctorMaterial
temperature = 'Ts'
rho = ${rho_s}
cp = ${cp_s}
block = 'c'
h = h_solid
[]
[]
[Executioner]
type = Transient
solve_type = NEWTON
petsc_options_iname = '-pc_type -ksp_gmres_restart -sub_pc_type -sub_pc_factor_shift_type'
petsc_options_value = 'asm 100 lu NONZERO'
line_search = 'none'
nl_rel_tol = 1e-8
dt = 10
end_time = 10
[]
[Outputs]
exodus = true
[] |
Beta Was this translation helpful? Give feedback.
Answered by
GiudGiud
Nov 23, 2024
Replies: 1 comment 3 replies
-
|
Hello When defining the entities, you chose the bounding boxes and they do not include some of the nodes. I did it quickly here |
Beta Was this translation helpful? Give feedback.
3 replies
-
|
Hi @GiudGiud, Thank you very much for your reply and the suggestion. Now I see what was wrong. Although I got a similar error with your suggested code, but I could run my model after making the larger bounding box larger for the circle surface (changing the last line before "$EndEntities") as below: In this case, wouldn't the larger physical bounding box affect the circle boundary in my model? I have set two surface entities to separate them into two blocks for different physics in MOOSE. So I need the elements outside of the circle boundary in the box to be still in the domain block having different physics from the ones in the circle. Also what could be the reason that this problem did not appear with the same mesh before I updated MOOSE? Could this be ignored if the node lies at the boundary for an older version? |
Beta Was this translation helpful? Give feedback.
-
I think you can check this by loading the mesh in moose then examining the output in Paraview
libmesh (which reads the .msh files) probably increased the input mesh checking. You can check the MOOSE newsletters for recent libmesh changes which might have affected you |
Beta Was this translation helpful? Give feedback.
-
|
Thanks a lot for the answers and for your time! It really helped. |
Beta Was this translation helpful? Give feedback.
Answer selected by
JuliaHLee
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hello
When defining the entities, you chose the bounding boxes
and they do not include some of the nodes.
If you want to use these nodes, you need to grow these bounding boxes
I did it quickly here