223031012
/
DAMASK_EICMD
mirror of https://github.com/achalhp/DAMASK_EICMD.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

faster version for large number of grains. now performing diffusion on a small window around each grain where window around each grain is obtained cheaply

This commit is contained in:
Pratheek Shanthraj 2013-06-27 16:57:14 +00:00
parent fd8d85896a
commit 4537720895
1 changed files with 51 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

63
processing/pre/geom_stretchInterfaces.py
View File

@ -49,13 +49,16 @@ The final geometry is assembled by selecting at each voxel that grain index for
)
parser.add_option('-t', '--time', dest='t', type='int', \
help='time for curvature flow [%default]')
help='time for curvature flow [%default]')
parser.add_option('-N', '--smooth', dest='N', type='int', \
help='number of steps for curvature flow [%default]')
parser.add_option('-b', '--black', dest='black', action='extend', type='string', \
help='indices of stationary microstructures', metavar='<LIST>')
parser.add_option('-2', '--twodimensional', dest='twoD', action='store_true', \
help='output geom file with two-dimensional data arrangement [%default]')
parser.set_defaults(t = 1)
parser.set_defaults(N = 1)
parser.set_defaults(black = [])
parser.set_defaults(twoD = False)
@ -149,20 +152,56 @@ for file in files:
i += 1
#--- initialize helper data -----------------------------------------------------------------------
winner = numpy.zeros(info['grid'],'i')
diffusedMax = numpy.zeros(info['grid'])
diffusionWindow = int(math.ceil(4.0*numpy.sqrt(options.t)))
for smoothIter in xrange(options.N):
extendedMicro = numpy.zeros(2*diffusionWindow+info['grid']).astype(int)
extendedMicro[:info['grid'][0],:info['grid'][1],:info['grid'][2]] = \
numpy.roll(numpy.roll(numpy.roll(microstructure,shift=diffusionWindow,axis=0),\
shift=diffusionWindow,axis=1),shift=diffusionWindow,axis=2)
extendedMicro[-info['grid'][0]:,:info['grid'][1],:info['grid'][2]] = \
numpy.roll(numpy.roll(numpy.roll(microstructure,shift=-diffusionWindow,axis=0),\
shift=diffusionWindow,axis=1),shift=diffusionWindow,axis=2)
extendedMicro[:info['grid'][0],-info['grid'][1]:,:info['grid'][2]] = \
numpy.roll(numpy.roll(numpy.roll(microstructure,shift=diffusionWindow,axis=0),\
shift=-diffusionWindow,axis=1),shift=diffusionWindow,axis=2)
extendedMicro[-info['grid'][0]:,-info['grid'][1]:,:info['grid'][2]] = \
numpy.roll(numpy.roll(numpy.roll(microstructure,shift=-diffusionWindow,axis=0),\
shift=-diffusionWindow,axis=1),shift=diffusionWindow,axis=2)
extendedMicro[:info['grid'][0],:info['grid'][1],-info['grid'][2]:] = \
numpy.roll(numpy.roll(numpy.roll(microstructure,shift=diffusionWindow,axis=0),\
shift=diffusionWindow,axis=1),shift=-diffusionWindow,axis=2)
extendedMicro[-info['grid'][0]:,:info['grid'][1],-info['grid'][2]:] = \
numpy.roll(numpy.roll(numpy.roll(microstructure,shift=-diffusionWindow,axis=0),\
shift=diffusionWindow,axis=1),shift=-diffusionWindow,axis=2)
extendedMicro[:info['grid'][0],-info['grid'][1]:,-info['grid'][2]:] = \
numpy.roll(numpy.roll(numpy.roll(microstructure,shift=diffusionWindow,axis=0),\
shift=-diffusionWindow,axis=1),shift=-diffusionWindow,axis=2)
extendedMicro[-info['grid'][0]:,-info['grid'][1]:,-info['grid'][2]:] = \
numpy.roll(numpy.roll(numpy.roll(microstructure,shift=-diffusionWindow,axis=0),\
shift=-diffusionWindow,axis=1),shift=-diffusionWindow,axis=2)
winner = numpy.zeros(extendedMicro.shape).astype(int)
winner[diffusionWindow:-diffusionWindow,diffusionWindow:-diffusionWindow,diffusionWindow:-diffusionWindow] =\
numpy.where(numpy.reshape(numpy.in1d(microstructure,options.black),microstructure.shape),\
microstructure,0)
diffusedMax = numpy.zeros(extendedMicro.shape)
boundingSlice = ndimage.measurements.find_objects(microstructure)
microList = set(numpy.unique(microstructure)).difference(set(options.black).union(set([0])))
#--- diffuse each grain separately ----------------------------------------------------------------
for theGrain in xrange(1,1+numpy.amax(microstructure)):
diffused = ndimage.filters.gaussian_filter((microstructure == theGrain).astype(float),\
{True:0.0,False:numpy.sqrt(options.t)}[theGrain in options.black],\
mode='wrap')
winner = numpy.where(diffused > diffusedMax, theGrain, winner)
diffusedMax = numpy.where(diffused > diffusedMax, diffused, diffusedMax)
microstructure = winner
for grain in microList:
xMin = boundingSlice[grain-1][0].start; xMax = boundingSlice[grain-1][0].stop + 2*diffusionWindow
yMin = boundingSlice[grain-1][1].start; yMax = boundingSlice[grain-1][1].stop + 2*diffusionWindow
zMin = boundingSlice[grain-1][2].start; zMax = boundingSlice[grain-1][2].stop + 2*diffusionWindow
diffused = ndimage.filters.gaussian_filter((extendedMicro[xMin:xMax,yMin:yMax,zMin:zMax] == grain).astype(float),\
numpy.sqrt(options.t))
isMax = diffused > diffusedMax[xMin:xMax,yMin:yMax,zMin:zMax]
winner[xMin:xMax,yMin:yMax,zMin:zMax][isMax] = grain
diffusedMax[xMin:xMax,yMin:yMax,zMin:zMax] = numpy.where(isMax,diffused,diffusedMax[xMin:xMax,yMin:yMax,zMin:zMax])
microstructure = winner[diffusionWindow:-diffusionWindow,diffusionWindow:-diffusionWindow,diffusionWindow:-diffusionWindow]
# --- assemble header -----------------------------------------------------------------------------
formatwidth = int(math.floor(math.log10(microstructure.max())+1))