[CIG-SHORT] fault traction perturbation

[Brad Aagaard]

On 11/18/2014 11:22 AM, Birendra jha wrote:
> Hi
>
> My question is with regard to the implementation of traction
> perturbation in dynamic faulting code. Traction perturbation is
> considered in the assembly of the residual (integrateResidual).
> However, it does not play any role anywhere else, especially in the
> solution correction steps. Is that correct? One of the correction
> step is "Insure fault traction is zero when opening." Is that still
> true when there is a positive traction perturbation?
>
> Thanks Birendra

The corrections to the solution to account for roundoff errors and other 
corner cases in FaultCohesiveDyn::constrainSolnSpace() does not need to 
include the traction perturbation. By including them in the residual, 
the Lagrange multipliers adjust automatically in the linear solves.

NOTE TO ALL USERS: When asking about specific routines or lines in the 
code, it really helps us if you specify line numbers in a file and 
either the source tarball version or a commit in the git repository.

For example,
https://github.com/geodynamics/pylith/blob/master/libsrc/pylith/faults/FaultCohesiveDyn.cc
Line 947

--or--

Line 949 in FaultCohesiveDyn.cc
branch: willic3/fix-plasticity
commit f34bab5c900421d8c52cbacc01a3da68dc85e663

Regards,
Brad