Strain rate solver for SSA

Extension of Elmer in computational glaciology
Post Reply
chenry
Posts: 13
Joined: 11 Nov 2020, 19:05
Antispam: Yes

Strain rate solver for SSA

Post by chenry »

Hi,

Is there a way to calculate strain rates (or even better, stresses) for simulations with the Shallow Shelf Approximation? The strain rate and stress solvers which have worked well for me for full Stokes simulations don’t work for SSA.

Code: Select all

Solver 6
  Equation = "Strain Rate"
  Procedure = "ElmerIceSolvers" "ComputeStrainRate"
! this is just a dummy, hence no output is needed
!-----------------------------------------------------------------------  
  Variable = -nooutput "Eij"
  Variable DOFs = 1

  Exported Variable 1 = "StrainRate"
  Exported Variable 1 DOFs = 5 !in 2D, 7 in 3D

! the name of the variable containing the flow solution (U,V,W,Pressure)
!-----------------------------------------------------------------------
  Flow Solver Name = String "SSAVelocity"
! the name of the strain-rate solution (default is 'StrainRate')
  StrainRate Variable Name = String "StrainRate"

  Linear System Solver = Direct
  Linear System Direct Method = umfpack
End
Thanks for any help!

Clara
kevinarden
Posts: 2435
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Strain rate solver for SSA

Post by kevinarden »

I don't think SSAVelocity is the name of the variable containing the flow solution
! the name of the variable containing the flow solution (U,V,W,Pressure)
!-----------------------------------------------------------------------
Flow Solver Name = String "SSAVelocity"



Equation = "SSA Velocity"
Procedure = File "ElmerIceSolvers" "SSASolver"
Variable = -nooutput String "varSSA"
Variable DOFs = 1

Exported Variable 1 = String "SSAFlow"
Exported Variable 1 DOFs = 4 ! 3 in 2D, 4 in 3D
chenry
Posts: 13
Joined: 11 Nov 2020, 19:05
Antispam: Yes

Re: Strain rate solver for SSA

Post by chenry »

Hi Kevin,

Thanks for the reply. Sorry, I forgot to mention that I am using "SSABasalSolver" rather than "SSASolver", which gives the velocity variable the name "SSAVelocity" rather than "SSAFlow". My understanding is that I need to use "SSABasalSolver" rather than "SSASolver" because I also need to solve for basal friction with my set up.

I have attached a sif file of my set up for solving the deviatoric stress and strain rate, as well as screenshots of the results. The results show a smooth velocity field, but incorrect dev. stress and strain rate solutions, with similar spatial patterns. Do you perhaps have an idea what might be going wrong in the solvers?

Thanks,

Clara
Attachments
Forward.sif
(10.92 KiB) Downloaded 13 times
Screenshot 2024-06-11 at 13.47.40.jpg
(219.15 KiB) Not downloaded yet
Screenshot 2024-06-11 at 13.46.16.jpg
(168 KiB) Not downloaded yet
kevinarden
Posts: 2435
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Strain rate solver for SSA

Post by kevinarden »

No I don't have a solution.
It bothers me that the flow solution is listed as U,V,W, and Pressure
but the SSAVelocity appears to be only U,V,W
! the name of the variable containing the flow solution (U,V,W,Pressure)
!-----------------------------------------------------------------------
Flow Solver Name = String "SSAVelocity"
chenry
Posts: 13
Joined: 11 Nov 2020, 19:05
Antispam: Yes

Re: Strain rate solver for SSA

Post by chenry »

I've done some more digging into the stress and strain rate fortran files and it seems that the pressure is only needed if solving for the Cauchy stress and seems to be read in separately to the flow solution anyway. My initial investigation suggests that the solvers are currently set up for 3D simulations, but not fully for 2D (3D simulations of stress and strain rate with full Stokes has worked perfectly for me previously). I'll update once I've had the chance to do some debugging.

-Clara
Post Reply