fatal element size

Clearly defined bug reports and their fixes
Post Reply
spacedout
Posts: 177
Joined: 30 Mar 2020, 23:27
Antispam: Yes

fatal element size

Post by spacedout »

Hi

I just happen to have a few linear wedge elements in my mesh (element code 706). Therefore Function ElementSize of file ElementDescription.F90

will round off 706/100 to 7 for the family integer in the select statement. However 7 is not one of the cases for that select resulting in a
default fatal call which prints:

ERROR:: ElementSize: Not implemented for elementtype

Not sure if it is a bug. Perhaps detJ (or size) is undefined for a wedge but I dont see why that should be so.

Thanks in advance for reading this post
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: fatal element size

Post by raback »

Thanx for reporting! This is now fixed in devel. -Peter
spacedout
Posts: 177
Joined: 30 Mar 2020, 23:27
Antispam: Yes

Re: fatal element size

Post by spacedout »

It works ! But later in the program execution I now get

ERROR:: ElementInfo: Bubbles for element: 706 are not implemented.
ERROR:: ElementInfo: Please use p-element basis instead.

which is printed out in file ElementDescription.F90 at the end of function ElementInfo in the section concerned with the generation of bubble basis functions (cases 6 and 7 are not implemented)

I dont know how tedious it is to add code for these cases. One easy way out of course is to set Bubbles to False in the SIF.
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: fatal element size

Post by raback »

Hi, can you share the solver settings that is causing this. I guess there is a way to use p-bubbles instead depending on solver. -Peter
spacedout
Posts: 177
Joined: 30 Mar 2020, 23:27
Antispam: Yes

Re: fatal element size

Post by spacedout »

Sure, no problem:

Solver 9
Equation = electron density
Procedure = "AdvectionDiffusion" "AdvectionDiffusionSolver"
Variable = Ne
Exec Solver = Always
Stabilize = False
Bubbles = True
Lumped Mass Matrix = False
Optimize Bandwidth = True
Steady State Convergence Tolerance = 1.0e-5
Nonlinear Post Solvers = 10
Nonlinear System Convergence Tolerance = 1.0e-8
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-8
Linear System Preconditioning = ILU1
Linear System ILUT Tolerance = 1.0e-3
Linear System Abort Not Converged = False
Linear System Residual Output = 1
Linear System Precondition Recompute = 1
End
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: fatal element size

Post by raback »

Hi

How would this work for you?

Code: Select all

Stabilize = True
Bubbles = False
-Peter
spacedout
Posts: 177
Joined: 30 Mar 2020, 23:27
Antispam: Yes

Re: fatal element size

Post by spacedout »

As expected there are no errors showing.

The only reason I was using

Stabilize = False
Bubbles = True

is because it is written in the Keywords section of the Advection Diffusion Equation under the Elmer Models Manual that

"It is more accurate and does not include any ad hoc terms. However, it may be computationally more expensive."

So I was aiming for more accuracy in spite of the fact there could be more number crunching involved. And I avoided ad hoc terms according to the above quote.
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: fatal element size

Post by raback »

Hi

I would not be too worried using stabilization. The "ad hoc" is maybe an overstatament. Basically the value for stabilization coefficient is well founded but the fact that the term is just single value implicitely assumes that aspect ratios in elements is close to one. The two methods usually give very similar results.

-Peter
Post Reply