Changeset 8397


Ignore:
Timestamp:
Apr 12, 2012, 9:00:49 PM (13 years ago)
Author:
davies
Message:

Adding comments

Location:
trunk/anuga_work/development/gareth/experimental/balanced_dev
Files:
2 edited

Legend:

Unmodified
Added
Removed
  • trunk/anuga_work/development/gareth/experimental/balanced_dev/swb2_domain.py

    r8385 r8397  
    106106        print '##########################################################################'
    107107        print '#'
    108         print '# Using shallow_water_balanced2 solver in experimental/balanced_dev/'
    109         print "#"
    110         print "# This solver is experimental - it may cause your computer to explode "
    111         print "#"
    112         print "# It is not expected to perform well in problems with very"
     108        print '# Using shallow_water_balanced2 solver in anuga_work/development/gareth/experimental/balanced_dev/'
     109        print "#"
     110        print "# This solver is experimental. Here are some tips on using it"
     111        print "#"
     112        print "# 1) When plotting outputs, I strongly suggest you examine centroid values, not vertex values"
     113        print "# , as the latter can be completely misleading near strong gradients in the flow. "
     114        print "# There is a util.py script in anuga_work/development/gareth/tests which might help you extract"
     115        print "# quantities at centroid values from sww files."
     116        print "# Note that to accuractely compute centroid values from sww files, the files need to store "
     117        print "# vertices uniquely. This makes for large sww files (3x), but is the price to pay for the right answer"
     118        print "# (unless we alter IO to allow centroids to be written to sww files, which would then affect"
     119        print "# ANUGA viewer as well -- I expect this would be lots of work)"
     120        print "#"
     121        print "# 2) In field scale applications (where the depth is typically > 1m), I suggest you set"
     122        print "# domain.minimum_allowed_height=0.01 (the default is 1.0e-3). With this setting, velocity"
     123        print "# seems to be very well behaved"
     124        print "#"
     125        print "# 3) This solver is not expected to perform well in problems with very"
    113126        print "# shallow water flowing down steep slopes (such that the stage_centroid_value "
    114         print "# is less than the maximum bed_edge_value on a given triangle) "
    115         print "#"
    116         print "# Also, it allows the stage_centroid_value to drop to the minimum bed_edge_value"
     127        print "# is less than the maximum bed_edge_value on a given triangle). However, analytical tests"
     128        print "# suggest it can do typical wetting/drying situations very well (parabolic oscillations test case) "
     129        print "#"
     130        print "# 4) This solver allows the stage_centroid_value to drop to the minimum bed_edge_value"
    117131        print "# on it's triangle. In other ANUGA versions (e.g. 1.2.1), the limit would be the"
    118132        print "# bed_centroid_value. This means that triangles store slightly more water than they are"
    119133        print "# typically interpreted to store, which might have significance in some applications."
    120134        print "#"
    121         print "# You will probably be able to tell if the above are causing you problems by convergence testing"
    122         print "#"
    123         print '# Note that many options in config.py have been overridden by the solver'
     135        print "# You will probably be able to tell this is causing you problems by convergence testing"
     136        print "#"
     137        print '# 5) Note that many options in config.py have been overridden by the solver -- I have '
     138        print '# deliberately attempted to get the solver to perform well with consistent values of '
     139        print '# these parameters -- so I would advise against changing them unless you at least check that '
     140        print '# it really does improve things'
    124141        print '#'
    125142        print '##########################################################################'
  • trunk/anuga_work/development/gareth/experimental/balanced_dev/swb2_domain_ext.c

    r8385 r8397  
    603603
    604604
    605   // Protect against initesimal and negative heights 
     605  // Protect against inifintesimal and negative heights 
    606606  if (maximum_allowed_speed < epsilon) {
    607607    for (k=0; k<N; k++) {
     
    630630  } else {
    631631   
    632     // Protect against initesimal and negative heights
     632    // Protect against infinitesimal and negative heights
    633633    for (k=0; k<N; k++) {
    634634      hc = wc[k] - zc[k];
Note: See TracChangeset for help on using the changeset viewer.