Input Domain

LBPM provides a flexible framework to ingest 3D image data. To illustrate the basic capabilities, this tutorial considers a quasi-2D flow cell. Source files for the example are included in the LBPM repository in the directory examples/DiscPack. A simple python code is included to set up the flow domain.

Based on LBPM convention, external boundary conditions are applied in the z-direction. This means that the domain should be set up so that the direction you want to set boundary conditions is aligned with the z-axis. For the quasi-2D example, a depth of 3 voxels is used for the x-direction. Based on LBPM internal data structures at least three voxels must be provided in each direction The specified domain decomposition must also observe this rule.

Image data is stored internally within LBPM as signed 8-bit binary data. This means that up to 256 labels can be provided in the input image. LBPM convention takes all non-positive labels to be immobile (treated as solid). In this example, the solid regions are assigned a value of 0. It is possible to provide up to 128 different labels for the solid. Also, note that python supports only the unsigned 8-bit datatype. For the unsigned data type, labels assigned values 128,...255 in python will correspond to labels -127,...-1 when read in as type signed char within LBPM.

  import numpy as np
  import matplotlib.pylab as plt
  import pandas as pd
  # Set the size of the domain
  Nx=3
  Ny=128
  Nz=128
  D=pd.read_csv("discs.csv",sep=" ")
  ID = np.ones(Nx*Ny*Nz,dtype='uint8')
  ID.shape = (Nz,Ny,Nx)
  # Set the solid labels
  for idx in range(len(D)):
     cx=D['cx'][idx] / dx
     cy=D['cy'][idx] /dx
     r=D['r'][idx] /dx
     for i in range(0,Nz):
        for j in range(0,Ny):
           if ( (cx-i)*(cx-i) + (cy-j)*(cy-j) < r*r ):
              ID[i,j,0] = 0
              ID[i,j,1] = 0
              ID[i,j,2] = 0
# write input file to disc
ID.tofile("discs_3x128x128.raw")