如何从FiPy(3D到2D)中的3D变量中提取平面 [英] How to extract a plane from a 3D variable in FiPy (3D to 2D)

查看:146
本文介绍了如何从FiPy(3D到2D)中的3D变量中提取平面的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

我在3D网格上有一个变量,我正在尝试切割一个计划.我很惊讶以前没有问过这个问题,这看起来是一个简单而常见的问题,但是我还没有找到任何好的方法.我将不胜感激.

I have a variable on a 3D mesh and I am trying to cut a plan. I am surprised this question hasn't been asked before, it looks an easy and common problem but I haven't found any good way. I would appreciate any advice.

假设我有一个平行六面体3x3x5,并且我正在尝试提取z平面.

Let's say that I have a parallelepiped 3x3x5 and that I am trying to extract a z-plane.

from fipy import *
from numpy import *
#Describes a 3x3x5 mesh
nx = 3
ny = 3
nz = 5

dx = 1
dy = 1
dz = 1

#Creates a 3D mesh and a 2D mesh to store the plane
mesh3D = Grid3D(dx, dy, dz, nx, ny, nz)
mesh2D = Grid2D(dx, dy, nx, ny)
#Defines the same variable, in 3D and in 2D
var2D = CellVariable(mesh = mesh2D)
var3D = CellVariable(mesh = mesh3D)

#Fills the 3D matrix
for i in xrange(nx*ny*nz*dx*dy*dz):
    var3D[i] = i

print var3D

输出:

[  0.   1.   2.   3.   4.   5.   6.   7.   8.   9.  10.  11.  12.  13.  14.  15.  16.  17.  18.  19.  20.  21.  22.  23.  24.  25.  26.  27.  28.  29.  30.  31.  32.  33.  34.  35.  36.  37.  38.  39.  40.  41.  42.  43.  44.]

3D变量看起来正确填充.

The 3D variable looks correctly filled.

首先,我尝试使用此链接中所述的方法 http://permalink.gmane.org/gmane.comp.python.fipy/1576

First, I tried to use the way described at this link http://permalink.gmane.org/gmane.comp.python.fipy/1576

CellVariable允许对传入的一组坐标进行插值 通过 call 方法(只需使用以下方法访问 call 方法 括号,就像在函数调用中一样).它返回一组值 对应于传入的每个坐标. order参数仅确定插值的顺序.

The call method of a CellVariable enables interpolation to a set of coordinates passed in via the call method (the call method is just accessed using parentheses like in a function call). It returns a set of values for corresponding to each of the coordinates that were passed in. The order argument just determines the order of the interpolation.

我不确定这个实际的工作原理,但是据我了解,这应该以0阶插值一个平面,因此它应该提取特定平面上的精确值.如果我错了,请纠正我.

I am not sure of how this actual works, but from what I understand this should interpolate a single plane with 0 order, so it should extract the exact value on a specific plane. Please correct me if I am wrong.

x3D, y3D, z3D = mesh3D.getCellCenters()
x2D, y2D =  mesh2D.getCellCenters()

for zcut in xrange(nz*dz):
    var2D.setValue(var3D((x2D, y2D, zcut * ones(var2D.getMesh().getNumberOfCells())), order=0))
    print "z-plane = %d" % zcut
    print var2D

raw_input("Press any key to close")

奇怪的是,它不起作用.偶数索引还可以,但奇数索引是连续平面的副本.

Strangely it doesn't work. The even indexes are ok but the odd ones are a copy of the contiguous planes.

z-plane = 0
[ 0.  1.  2.  3.  4.  5.  6.  7.  8.]
z-plane = 1
[ 0.  1.  2.  3.  4.  5.  6.  7.  8.]
z-plane = 2
[ 18.  19.  20.  21.  22.  23.  24.  25.  26.]
z-plane = 3
[ 18.  19.  20.  21.  22.  23.  24.  25.  26.]
z-plane = 4
[ 36.  37.  38.  39.  40.  41.  42.  43.  44.]

我认为某处肯定有一个愚蠢的错误,但我不知道.有什么主意吗?

I think there must be a stupid mistake somewhere, but I have no clue. Any idea?

推荐答案

像元中心位于z = 0.5、1.5、2.5等位置,因此FiPy最好找到最接近z = 0的像元, 1,2,...

The cell centers are located at z=0.5, 1.5, 2.5, ... so FiPy is doing it's best to find the nearest cells to z=0, 1, 2, ...

尝试

var2D.setValue(var3D((x2D, y2D, 
                      zcut * ones(var2D.mesh.numberOfCells) + dx/2.), order=0))

这篇关于如何从FiPy(3D到2D)中的3D变量中提取平面的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆