Homec4science
Diffusion GooseFEM (main) Browse

rGOOSEFEM/docs/examples/MeshTri3/ReTri_random.py
8949d67c7af9main

Last Change
Raw File

ReTri_random.py
View Options

import numpy as np
import GooseFEM
# define a mesh, and extract the relevant data
mesh = GooseFEM.Mesh.Tri3.Regular(10,10)
coor = mesh.coor()
conn = mesh.conn()
nnode = coor.shape[0]
nelem = conn.shape[0]
per = mesh.nodesPeriodic()
# define a random perturbation of the nodal coordinates
delta = np.random.random(coor.shape)
delta[:,0] *= .1
delta[:,1] *= .1
# make periodic
delta[per[:,1],0] = delta[per[:,0],0]
delta[per[:,1],1] = delta[per[:,0],1]
# copy the coordinates and connectivity (for plotting)
newcoor = np.array(coor,copy=True)
newconn = np.array(conn,copy=True)
# add the perturbation incrementally, compute the new connectivity each time
for i in range(4):
newcoor += delta/4.
newconn = GooseFEM.Mesh.Tri3.retriangulate(newcoor,newconn)
# ==================================================================================================
# define colors to clearly show the difference (and element overlaps)
orig = np.zeros(( nelem ))
diff = np.zeros(( nelem ))
# find the differences
for i in range(nelem):
if not np.all( conn[i,:] == newconn[i,:] ):
diff[i] = 1
# define shifts to show periodicity
DX = np.zeros( coor.shape )
DY = np.zeros( coor.shape )
DX[:,0] = +1.
DY[:,1] = +1.
# --------------------------------------------------------------------------------------------------
import matplotlib.pylab as plt
import goosempl as gplt
plt.style.use(['goose','goose-latex'])
fig = plt.figure(figsize=(18,6))
fig.set_tight_layout(True)
# --------------------------------------------------------------------------------------------------
ax = fig.add_subplot(1,3,1)
gplt.patch(coor=coor,conn=conn,cindex=orig,clim=[-0.5,1.5],cmap='RdBu_r',alpha=.5)
gplt.patch(coor=coor-DX ,conn=conn)
gplt.patch(coor=coor+DX ,conn=conn)
gplt.patch(coor=coor+DX-DY,conn=conn)
gplt.patch(coor=coor+DX+DY,conn=conn)
gplt.patch(coor=coor-DX-DY,conn=conn)
gplt.patch(coor=coor-DX+DY,conn=conn)
gplt.patch(coor=coor -DY,conn=conn)
gplt.patch(coor=coor +DY,conn=conn)
plt.xlim([ -0.15 , 1.2 ])
plt.ylim([ -0.15 , 1.2 ])
ax.set_aspect('equal')
# --------------------------------------------------------------------------------------------------
ax = fig.add_subplot(1,3,2)
gplt.patch(coor=newcoor,conn=conn,cindex=orig,clim=[-0.5,1.5],cmap='RdBu_r',alpha=.5)
gplt.patch(coor=newcoor-DX ,conn=conn)
gplt.patch(coor=newcoor+DX ,conn=conn)
gplt.patch(coor=newcoor+DX-DY,conn=conn)
gplt.patch(coor=newcoor+DX+DY,conn=conn)
gplt.patch(coor=newcoor-DX-DY,conn=conn)
gplt.patch(coor=newcoor-DX+DY,conn=conn)
gplt.patch(coor=newcoor -DY,conn=conn)
gplt.patch(coor=newcoor +DY,conn=conn)
plt.xlim([ -0.15 , 1.2 ])
plt.ylim([ -0.15 , 1.2 ])
ax.set_aspect('equal')
# --------------------------------------------------------------------------------------------------
ax = fig.add_subplot(1,3,3)
gplt.patch(coor=newcoor,conn=newconn,cindex=diff,clim=[-0.5,1.5],cmap='RdBu_r',alpha=.5)
gplt.patch(coor=newcoor-DX ,conn=newconn)
gplt.patch(coor=newcoor+DX ,conn=newconn)
gplt.patch(coor=newcoor+DX-DY,conn=newconn)
gplt.patch(coor=newcoor+DX+DY,conn=newconn)
gplt.patch(coor=newcoor-DX-DY,conn=newconn)
gplt.patch(coor=newcoor-DX+DY,conn=newconn)
gplt.patch(coor=newcoor -DY,conn=newconn)
gplt.patch(coor=newcoor +DY,conn=newconn)
plt.xlim([ -0.15 , 1.2 ])
plt.ylim([ -0.15 , 1.2 ])
ax.set_aspect('equal')
# --------------------------------------------------------------------------------------------------
plt.savefig('ReTri_random.svg')
plt.show()
c4science · Help