Changes between Version 2 and Version 3 of UnderstandingErrorsQuestions


Ignore:
Timestamp:
Nov 13, 2008, 2:06:02 PM (15 years ago)
Author:
ole
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • UnderstandingErrorsQuestions

    v2 v3  
    99
    1010ANUGA uses variable timestepping to guarantee that information never skips a triangle. This is described in the user manual in the section outlining the Finite Volume Method. If velocities for one reason or another grow without bound, the automatic timestepping will eventually cause ANUGA to grind to a halt with an error message like:
    11 
     11{{{
    1212Exception: WARNING: Too small timestep 0.0000000000000000 reached even after 50 steps of 1 order scheme
     13}}}
     14One typical reason for this is numerical instability which although rare in ANUGA can occur under special circumstances such as those associated with the transmissive boundary conditions mentioned in the manual. Another typical reason is excessive values in either boundary conditions or forcing terms.
    1315
    14 One typical reason for this is numerical instability which although rare in ANUGA can occur under special circumstances such as those associated with the transmissive boundary conditions mentioned in the manual. Another typical reason is excessive values in either boundary conditions or forcing terms.
     16So if this error is encountered, the first action is replace any transmissive boundary conditions with either Dirichlet or Time boundary objects. If that doesn't work, the next action is to reconsider values introduced by all boundary conditions and forcing terms.