[CIG-SEISMO] Unresolved issues in specfem2d

[Dimitri Komatitsch]

Hi Alexis, Hi all,

Thanks a lot. Very useful!

I will have a look at 4 and 5. (in two or three weeks)

4 could be classical Hourglass-like modes (see 
http://www.uni-magdeburg.de/ifme/zeitschrift_tm/2014_Heft2/01_Schmicker_et_al.pdf), 
i.e. not a bug but a property of finite element methods. Bernhard 
Schuberth had noticed that in 2003 already (in the 3D code).

5 could be related to the viscoelastic memory variables being in L^2 
rather than H^1, if so, same thing, not a bug but a "feature".

Zhinan, could your student have a look at 1, 2 and 3?

Thanks!
Best wishes,
Dimitri.


On 03/09/2018 03:06 PM, Alexis Bottero wrote:
> Hello everybody,
> I just commited a new folder called unresolvedBugs in specfem2d/EXAMPLES 
> in order to gather all unresolved issues. The examples there are 
> supposed to be simple, working, not containing useless features and have 
> a minimum computational cost.
> I think that it can be useful for the community.
> 
> For the moment I have added five cases that I have detected during my PhD.
> 
> 1. bugPMLs_1 : Strong spurious reflections from the PMLs in some 
> particular cases.
> 
> 2. bugPMLs_2 : This bug (the worse by far) occurs in simulations with 
> many time steps (>100000) and viscoelastic media with PMLs. Even if the 
> PMLs are not in contact with the viscoelastic part!
> 
> 3. bugPMLs_3 : Small bug when using PMLs with the internal mesher in 
> some particular configurations.
> 
> 4. bugSource : Strange (small) unphysical dispersion in acoustic domains.
> 
> 5. bugViscoelasticity : This shows a gradual, very slow explosion in 
> viscoelastic media. This issue is probably not a "bug" strictly speaking 
> but may arise from the inconsistency of memory variables between two 
> adjacent elements. The real problem occurs when considering also 
> acoustic media in the same run. In that case the instability occurs faster.
> 
> Do not hesitate to take a look at these examples and add yours!
> 
> Thank you very much!
> 
> Cheers
> 
> Alexis Bottero, PhD student
> Tel: +0033 6 95 17 00 97
> Waves and imaging
> Laboratoire de Mécanique et d'Acoustique de Marseille
> 
> 
> _______________________________________________
> CIG-SEISMO mailing list
> CIG-SEISMO at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo
> 

-- 
Dimitri Komatitsch, CNRS Research Director (DR CNRS)
Laboratory of Mechanics and Acoustics, Marseille, France
http://komatitsch.free.fr