Custom Use

Introduction

nsfds2 can also be used as a classical Python package. It provides several main objects to perform numerical simulations :

init package povides nsfds2.init.config.CfgSetup that parses all simulation parameters from ~/.nsfds2/nsfds2.conf
fdtd module provides nsfds2.fdtd.FDTD that runs the simulation
utils package provides in particular nsfds2.utils.graphics module useful for post-treatment

The following example gives the general philosophy to use nsfds2:

import matplotlib.pyplot as plt
from nsfds2.init import CfgSetup
from nsfds2.fdtd import FDTD
from fdgrid.mesh import Mesh

# Read simulation parameter in config file ~/nsfds2/nsfds2.conf
cfg = CfgSetup()    # or cfg = CfgSetup('path_to_configfile.conf')

# Define the mesh
msh = Mesh((cfg.nx, cfg.nz), (cfg.dx, cfg.dz), origin=(cfg.ix0, cfg.iz0), obstacles=cfg.obstacles)

# Create simulation
fdtd = FDTD(msh, cfg)
fdtd.run()

# Figures
plt.figure()
plt.imshow(fdtd.fld.p)
plt.show()

Once the simulation is finished, you can access the acoustic field at the last iteration through the fdtd.fld object using following attributes.

fdtd.fld attributes	description
r	density
ru	product of density and x component of velocity
rv	product of density and z component of velocity
re	product of density and energy

hdf5 files

If save option is selected, hdf5 are created. They contain the following variables.

var	variable
r_init	initial density
ru_init	initial product of density and x component of velocity
rv_init	initial product of density and z component of velocity
re_init	initial product of density and energy
r_itX	density
ru_itX	product of density and x component of velocity
rv_itX	product of density and z component of velocity
re_itX	product of density and energy
x	x-grid
z	z-grid
probe_locations	coordinates of probes
probe_values	pressure at probe locations
obstacles	coordinates of obstacles

The r_itX, ru_itX, rv_itX, and re_itX quantities gather the field at each time iteration (2d array).

To access the acoustic pressure, one can use:

import numpy as np
from ofdlib2 import fdtd
...
p  = np.empty_like(rho)
fdtd.p(p, r, ru, rv, re, gamma)

The variable p needs to be initialize first. Then, it is used as an input argument of fdtd.p

Note that in curvilinear coordinates, there are following sets of coordinates :

var	variable
xn	numerical x-grid
zn	numerical z-grid
xp	physical x-grid
zp	physical z-grid
J	Jacobian matrix of transformation

cfg pickle files

A cfg file is also created for each simulation gathering its configuration. It contains the configuration object relative to the simulation:

import pickle

with open(filename, 'rb') as _file:
    cfg = pickle.load(_file)

This object contains in particular :

cfg.dx, cfg.dz, cfg.dt : spatial and time steps
cfg.nx, cfg.nz, cfg.nt, cfg.ns : Number of points (spatial and temporal)
cfg.p0, cfg.rho0, cfg.T0, cfg.c0, cfg.gamma, cfg.prandtl, cfg.mu : Thermophysical parameters
… and many other parameters.