代码之家 › 专栏 › 技术社区 › Tom de Geus

如何读取有限元网格并将其视为非结构化网格

paraview

0

Tom de Geus · 技术社区 · 8 年前

我知道如何做一些简单的事情:

from paraview.simple import *

Sphere()
Show()
Render()

但是,如何制作网格?让我们跳过HDF5部分,重点关注这个非常简单的二维网格,包括二维四边形:

from paraview.simple import *
import numpy as np

coor = np.array([
    [ 0.0 , 0.0 ],
    [ 1.0 , 0.0 ],
    [ 2.0 , 0.0 ],
    [ 0.0 , 1.0 ],
    [ 1.0 , 1.0 ],
    [ 2.0 , 1.0 ],
])

conn = np.array([
    [ 0 , 1 , 4 , 3 ],
    [ 1 , 2 , 5 , 4 ],
])

一种可能的方法是建立一个 vtkUnstructuredGrid .但问题是我不知道该怎么办。一、 e.我如何告诉ParaView使用它?

from paraview import vtk

points = vtk.vtkPoints()

for i,(x,y) in enumerate(coor):
    points.InsertNextPoint(x,y,0.0)

grid = vtk.vtkUnstructuredGrid()

for el in conn:
    cell = vtk.vtkQuad()
    for i,ver in enumerate(el): 
        cell.GetPointIds().SetId(i,ver)
    grid.InsertNextCell(cell.GetCellType(),cell.GetPointIds())

grid.SetPoints(points)

理解如何包含单元数据和点数据将非常有帮助,但我可能可以从这个问题的解决方案开始解决这个问题。

2 回复 | 直到 8 年前

1

2

Cory Quammen 8 年前

在较低的级别上,可以使用Python定义VTK的操作。 Programmable Filter 和 Programmable Source 可编程源 paraview.simple paraview.simple

可编程源 ,设置其 Output Data Set Type 到 vtkUnstructuredGrid

import numpy as np

coor = np.array([
    [ 0.0 , 0.0 ],
    [ 1.0 , 0.0 ],
    [ 2.0 , 0.0 ],
    [ 0.0 , 1.0 ],
    [ 1.0 , 1.0 ],
    [ 2.0 , 1.0 ],
])

conn = np.array([
    [ 0 , 1 , 4 , 3 ],
    [ 1 , 2 , 5 , 4 ],
])

import vtk

grid = self.GetOutput()

points = vtk.vtkPoints()
for i,(x,y) in enumerate(coor):
    points.InsertNextPoint(x,y,0.0)
grid.SetPoints(points)

grid.Allocate()
for el in conn:
    cell = vtk.vtkQuad()
    for i,ver in enumerate(el): 
          cell.GetPointIds().InsertId(i,ver)
    grid.InsertNextCell(cell.GetCellType(),cell.GetPointIds())

2

0

Tom de Geus 8 年前

请相信@CoryQuammen,因为这确实是他的解决方案

完整脚本,包括一些单元格和点数据,供将来参考

from paraview.simple import *

paraview.simple._DisableFirstRenderCameraReset()

# create a new 'Programmable Source'
mesh = ProgrammableSource()
mesh.OutputDataSetType = 'vtkUnstructuredGrid'
mesh.ScriptRequestInformation = ''
mesh.PythonPath = ''
mesh.Script = '''
import numpy as np

coor = np.array([
    [ 0.0 , 0.0 ],
    [ 1.0 , 0.0 ],
    [ 2.0 , 0.0 ],
    [ 0.0 , 1.0 ],
    [ 1.0 , 1.0 ],
    [ 2.0 , 1.0 ],
])

conn = np.array([
    [ 0 , 1 , 4 , 3 ],
    [ 1 , 2 , 5 , 4 ],
])

celldata = [
  1.0 ,
  2.0
]

normals = np.array([
  [ -1.0 , -1.0 ],
  [  0.0 , -1.0 ],
  [ +1.0 , -1.0 ],
  [ -1.0 , +1.0 ],
  [  0.0 , +1.0 ],
  [ +1.0 , +1.0 ],
])

normals /= np.tile(np.sqrt(np.sum(normals**2.,axis=1)).reshape(-1,1),(1,2))

import vtk

grid = self.GetOutput()

points = vtk.vtkPoints()
for i,(x,y) in enumerate(coor):
    points.InsertNextPoint(x,y,0.0)
grid.SetPoints(points)

grid.Allocate()
for el in conn:
    cell = vtk.vtkQuad()
    for i,ver in enumerate(el):
          cell.GetPointIds().InsertId(i,ver)
    grid.InsertNextCell(cell.GetCellType(),cell.GetPointIds())

data = vtk.vtkDoubleArray()
data.SetName("Example data")
for i in celldata:
  data.InsertNextValue(i)

grid.GetCellData().AddArray(data)

data = vtk.vtkDoubleArray()
data.SetNumberOfComponents(3)
data.SetName("Normals")
for i in normals:
  data.InsertNextTuple([i[0],i[1],0.0])

grid.GetPointData().AddArray(data)
'''

# show data from mesh
Mesh = Show(mesh)

Render()