search

SeisSol / SeisSol

A scientific software for the numerical simulation of seismic wave phenomena and earthquake dynamics
http://www.seissol.org
BSD 3-Clause "New" or "Revised" License
265 stars 114 forks source link

Numerical artifacts at the bottom of a fault zone? #352

Open NicoSchlw opened 3 years ago

NicoSchlw commented 3 years ago

We observe some weird slipping behavior of some elements at the bottom of a fault zone (thin layer of drastically reduced seismic velocities and elastic moduli around the shallow part of the fault).

I wrote a script that calculates the rise time (time the slip rate is above a certain threshold) and found that these elements are slipping almost the entire simulation time (100s): grafik

The produced absolute slip is very low, but is visible when the colorbar is saturated: grafik

The fault zone leads to trapped waves, and therefore to strong stress concentrations. Dynamic triggering of these fault parts would not be something unexpected, but the distribution of elements slipping the entire simulation time, next to elements that do not slip at all, does not look physical.

We used the seissol branch thomas/RS_positive_normal_stress and ran it on SuperMuc. The mesh resolution on fault is 75m and we used order 5.

palgunadi1993 commented 2 years ago

Hi, could this be related to creeping associated with low traction? Did you solve the problem?

NicoSchlw commented 2 years ago

Hi, the slipping elements are within a strongly velocity-weakening regime, there should be no major creeping. I did not solve the problem, but I assume the observed slip rates are too low to have an impact on the results.

sebwolf-de commented 2 years ago

Hi, could you try the dr/cpp branch or the master branch after #604