Python 3D FDTD Simulator
A 3D electromagnetic FDTD simulator written in Python. The FDTD simulator has an optional PyTorch backend, enabling FDTD simulations on a GPU.
Docs
Examples
00. Quick Start
Imports
the fdtd library is simply imported as follows:
[1]:
import fdtd
Setting the backend
the fdtd library allows for setting a backend. There exist a Numpy backend and several PyTorch backends. The available backends are: - "numpy"
(defaults to float64 arrays) - "torch"
(defaults to float64 tensors) - "torch.float32"
- "torch.float64"
- "torch.cuda"
(defaults to float64 tensors) - "torch.cuda.float32"
- "torch.cuda.float64"
In general, the "numpy"
backend is preferred for standard CPU calculations with "float64"
precision. In general, "float64"
precision is always preferred for FDTD simulations, however, "float32"
might give a significant performance boost.
The "cuda"
backends are only available for computers with a GPU.
[2]:
fdtd.set_backend("numpy")
The FDTD-grid
The FDTD grid defines the simulation region.
# signature
fdtd.Grid(
shape: Tuple[Number, Number, Number],
grid_spacing: float = 155e-9,
permittivity: float = 1.0,
permeability: float = 1.0,
courant_number: float = None,
)
A grid is defined by its shape
, which is just a 3D tuple of Number
-types (integers or floats). If the shape is given in floats, it denotes the width, height and length of the grid in meters. If the shape is given in integers, it denotes the width, height and length of the grid in terms of the grid_spacing
. Internally, these numbers will be translated to three integers: grid.Nx
, grid.Ny
and grid.Nz
.
A grid_spacing
can be given. For stability reasons, it is recommended to choose a grid spacing that is at least 10 times smaller than the smallest wavelength in the grid. This means that for a grid containing a source with wavelength 1550nm
and a material with refractive index of 3.1
, the recommended minimum grid_spacing
turns out to be 50nm
For the permittivity
and permeability
floats or arrays with the following shapes
(grid.Nx, grid.Ny, grid.Nz)
or
(grid.Nx, grid.Ny, grid.Nz, 1)
or
(grid.Nx, grid.Ny, grid.Nz, 3)
are expected. In the last case, the shape implies the possibility for different permittivity for each of the major axes (so-called uniaxial or biaxial materials). Internally, these variables will be converted (for performance reasons) to their inverses grid.inverse_permittivity
array and a grid.inverse_permeability
array of shape (grid.Nx, grid.Ny, grid.Nz, 3)
. It is possible to change those arrays after making the grid.
Finally, the courant_number
of the grid determines the relation between the time_step
of the simulation and the grid_spacing
of the grid. If not given, it is chosen to be the maximum number allowed by the Courant-Friedrichs-Lewy Condition: 1
for 1D
simulations, 1/√2
for 2D
simulations and 1/√3
for 3D
simulations (the dimensionality will be derived by the shape of the grid). For stability
reasons, it is recommended not to change this value.
[3]:
grid = fdtd.Grid(
shape = (25e-6, 15e-6, 1), # 25um x 15um x 1 (grid_spacing) --> 2D FDTD
)
print(grid)
Grid(shape=(161,97,1), grid_spacing=1.55e-07, courant_number=0.70)
Adding an object to the grid
An other option to locally change the permittivity
or permeability
in the grid is to add an Object
to the grid.
# signature
fdtd.Object(
permittivity: Tensorlike,
name: str = None
)
An object defines a part of the grid with modified update equations, allowing to introduce for example absorbing materials or biaxial materials for which mixing between the axes are present through Pockels coefficients
or many more. In this case we’ll make an object with a different permittivity
than the grid it is in.
Just like for the grid, the Object
expects a permittivity
to be a floats or an array of the following possible shapes
(obj.Nx, obj.Ny, obj.Nz)
or
(obj.Nx, obj.Ny, obj.Nz, 1)
or
(obj.Nx, obj.Ny, obj.Nz, 3)
Note that the values obj.Nx
, obj.Ny
and obj.Nz
are not given to the object constructor. They are in stead derived from its placing in the grid:
[4]:
grid[11:32, 30:84, 0] = fdtd.Object(permittivity=1.7**2, name="object")
Several things happen here. First of all, the object is given the space [11:32, 30:84, 0]
in the grid. Because it is given this space, the object’s Nx
, Ny
and Nz
are automatically set. Furthermore, by supplying a name to the object, this name will become available in the grid:
[5]:
print(grid.object)
Object(name='object')
@ x=11:32, y=30:84, z=0:1
We can add a second object to the grid:
[6]:
grid[13e-6:18e-6, 5e-6:8e-6, 0] = fdtd.Object(permittivity=1.5**2)
Here we chose to slice the grid with floating point numbers, which will be replaced by integer Nx
, Ny
and Nz
during the registration of the object. Since we didnt give the object a name, the object won’t be available to us as an attribute of the grid. However, it is still available to us via the grid.objects
list:
[7]:
print(grid.objects)
[Object(name='object'), Object(name=None)]
This list stores all objects (i.e. of type fdtd.Object
) in the order that they were added to the grid.
Adding a source to the grid
Similarly as to adding an object to the grid, an fdtd.LineSource
can also be added:
# signature
fdtd.LineSource(
period: Number = 15, # timesteps or seconds
amplitude: float = 1.0,
phase_shift: float = 0.0,
name: str = None,
)
Similarly to an fdtd.Object
, an fdtd.Source
size is defined by its placement on the grid:
[8]:
grid[7.5e-6:8.0e-6, 11.8e-6:13.0e-6, 0] = fdtd.LineSource(
period = 1550e-9 / (3e8), name="source"
)
However, it is important to note that in this case we are adding a LineSource
, i.e. the source spans the diagonal of the cube defined by the slices. Internally, these slices will be converted into lists to ensure the expected behavior:
[9]:
print(grid.source)
LineSource(period=14, amplitude=1.0, phase_shift=0.0, name='source')
@ x=[48, ... , 51], y=[76, ... , 83], z=[0, ... , 0]
Note that one could have also supplied lists to index the grid in the first place. This feature could be useful to create a LineSource
of arbitrary shape.
Adding a detector to the grid
Adding a detector to the grid works the same as adding a source
# signature
fdtd.LineDetector(
name=None
)
[10]:
grid[12e-6, :, 0] = fdtd.LineDetector(name="detector")
[11]:
print(grid.detector)
LineDetector(name='detector')
@ x=[77, ... , 77], y=[0, ... , 96], z=[0, ... , 0]
Adding grid boundaries
Although, having an object, source and detector to simulate is in principle enough to perform an FDTD simulation, One also needs to define a grid boundary to prevent the fields to be reflected. One of those boundaries that can be added to the grid is a Perfectly Matched Layer or PML
. These are basically absorbing boundaries.
fdtd.PML(
a: float = 1e-8, # stability factor
name: str = None
)
[12]:
# x boundaries
# grid[0, :, :] = fdtd.PeriodicBoundary(name="xbounds")
grid[0:10, :, :] = fdtd.PML(name="pml_xlow")
grid[-10:, :, :] = fdtd.PML(name="pml_xhigh")
# y boundaries
# grid[:, 0, :] = fdtd.PeriodicBoundary(name="ybounds")
grid[:, 0:10, :] = fdtd.PML(name="pml_ylow")
grid[:, -10:, :] = fdtd.PML(name="pml_yhigh")
Grid summary
A simple summary of the grid can be shown by printing out the grid:
[13]:
print(grid)
Grid(shape=(161,97,1), grid_spacing=1.55e-07, courant_number=0.70)
sources:
LineSource(period=14, amplitude=1.0, phase_shift=0.0, name='source')
@ x=[48, ... , 51], y=[76, ... , 83], z=[0, ... , 0]
detectors:
LineDetector(name='detector')
@ x=[77, ... , 77], y=[0, ... , 96], z=[0, ... , 0]
boundaries:
PML(name='pml_xlow')
@ x=0:10, y=:, z=:
PML(name='pml_xhigh')
@ x=-10:, y=:, z=:
PML(name='pml_ylow')
@ x=:, y=0:10, z=:
PML(name='pml_yhigh')
@ x=:, y=-10:, z=:
objects:
Object(name='object')
@ x=11:32, y=30:84, z=0:1
Object(name=None)
@ x=84:116, y=32:52, z=0:1
Running a simulation
Running a simulation is as simple as using the grid.run
method.
grid.run(
total_time: Number,
progress_bar: bool = True
)
Just like for the the lengths in the grid, the total_time
of the simulation can be specified as an integer (number of time_steps
) or as a float (in seconds).
[14]:
grid.run(total_time=100)
100%|██████████| 100/100 [00:00<00:00, 526.29it/s]
Grid visualization
Let’s visualize the grid. This can be done with the grid.visualize
method:
# signature
grid.visualize(
grid,
x=None,
y=None,
z=None,
cmap="Blues",
pbcolor="C3",
pmlcolor=(0, 0, 0, 0.1),
objcolor=(1, 0, 0, 0.1),
srccolor="C0",
detcolor="C2",
show=True,
)
This method will by default visualize all objects in the grid, as well as the current field intensity at a certain x
, y
OR z
-plane. By setting show=False
, one can disable the immediate visualization of matplotlib.
[15]:
grid.visualize(z=0, show=False)
import matplotlib.pyplot as plt

01. Basic Example
A simple example on how to use the FDTD Library
Imports
[1]:
import matplotlib.pyplot as plt
import fdtd
import fdtd.backend as bd
Set Backend
[2]:
fdtd.set_backend("numpy")
Constants
[3]:
WAVELENGTH = 1550e-9
SPEED_LIGHT: float = 299_792_458.0 # [m/s] speed of light
Simulation
create FDTD Grid
[4]:
grid = fdtd.Grid(
(2.5e-5, 1.5e-5, 1),
grid_spacing=0.1 * WAVELENGTH,
permittivity=1.0,
permeability=1.0,
)
boundaries
[5]:
# grid[0, :, :] = fdtd.PeriodicBoundary(name="xbounds")
grid[0:10, :, :] = fdtd.PML(name="pml_xlow")
grid[-10:, :, :] = fdtd.PML(name="pml_xhigh")
# grid[:, 0, :] = fdtd.PeriodicBoundary(name="ybounds")
grid[:, 0:10, :] = fdtd.PML(name="pml_ylow")
grid[:, -10:, :] = fdtd.PML(name="pml_yhigh")
grid[:, :, 0] = fdtd.PeriodicBoundary(name="zbounds")
sources
[6]:
grid[50:55, 70:75, 0] = fdtd.LineSource(
period=WAVELENGTH / SPEED_LIGHT, name="linesource"
)
grid[100, 60, 0] = fdtd.PointSource(
period=WAVELENGTH / SPEED_LIGHT, name="pointsource",
)
detectors
[7]:
grid[12e-6, :, 0] = fdtd.LineDetector(name="detector")
objects
[8]:
grid[11:32, 30:84, 0:1] = fdtd.AnisotropicObject(permittivity=2.5, name="object")
Run simulation
[9]:
grid.run(50, progress_bar=False)
Visualization
[10]:
fig, axes = plt.subplots(2, 3, squeeze=False)
titles = ["Ex: xy", "Ey: xy", "Ez: xy", "Hx: xy", "Hy: xy", "Hz: xy"]
fields = bd.stack(
[
grid.E[:, :, 0, 0],
grid.E[:, :, 0, 1],
grid.E[:, :, 0, 2],
grid.H[:, :, 0, 0],
grid.H[:, :, 0, 1],
grid.H[:, :, 0, 2],
]
)
m = max(abs(fields.min().item()), abs(fields.max().item()))
for ax, field, title in zip(axes.ravel(), fields, titles):
ax.set_axis_off()
ax.set_title(title)
ax.imshow(bd.numpy(field), vmin=-m, vmax=m, cmap="RdBu")
plt.show()

[11]:
plt.figure()
grid.visualize(z=0)

02. Absorbing Object
A simple example using the AbsorbingObject
Imports
[1]:
import matplotlib.pyplot as plt
import fdtd
fdtd.set_backend("numpy")
Constants
[2]:
WAVELENGTH = 1550e-9
SPEED_LIGHT: float = 299_792_458.0 # [m/s] speed of light
Grid setup
[3]:
# create FDTD Grid
grid = fdtd.Grid(
(1.5e-5, 1.5e-5, 1), # 2D grid
grid_spacing=0.1 * WAVELENGTH,
permittivity=2.5, # same as object
)
# sources
grid[15, :] = fdtd.LineSource(period=WAVELENGTH / SPEED_LIGHT, name="source")
# detectors
grid[-15, :, 0] = fdtd.LineDetector(name="detector")
# x boundaries
# grid[0, :, :] = fdtd.PeriodicBoundary(name="xbounds")
grid[0:10, :, :] = fdtd.PML(name="pml_xlow")
grid[-10:, :, :] = fdtd.PML(name="pml_xhigh")
# y boundaries
# grid[:, 0, :] = fdtd.PeriodicBoundary(name="ybounds")
grid[:, 0:10, :] = fdtd.PML(name="pml_ylow")
grid[:, -10:, :] = fdtd.PML(name="pml_yhigh")
# The absorbing object fills the whole grid
grid[10:-10, 10:-10, :] = fdtd.AbsorbingObject(
permittivity=2.5, conductivity=1e-6, name="absorbin_object"
)
Simulation without absorption:
[4]:
grid.run(250, progress_bar=False)
Visualization
[5]:
grid.visualize(z=0)

03. Objects of arbitrary shape
Imports
[1]:
import fdtd
import numpy as np
import matplotlib.pyplot as plt
fdtd.set_backend("numpy")
Grid Setup
[2]:
grid = fdtd.Grid(
shape = (300, 300, 1), # 25um x 15um x 1 (grid_spacing) --> 2D FDTD
grid_spacing = 1e-7,
permittivity = 1,
)
grid[50:250, 50, 0] = fdtd.LineSource(
period = 1550e-9 / (3e8), name="source"
)
grid[50:250, 250, 0] = fdtd.LineDetector(name="detector")
grid[0:10, :, :] = fdtd.PML(name="pml_xlow")
grid[-10:, :, :] = fdtd.PML(name="pml_xhigh")
grid[:, 0:10, :] = fdtd.PML(name="pml_ylow")
grid[:, -10:, :] = fdtd.PML(name="pml_yhigh")
Circular Object
[3]:
refractive_index = 1.7
x = y = np.linspace(-1,1,100)
X, Y = np.meshgrid(x, y)
circle_mask = X**2 + Y**2 < 1
permittivity = np.ones((100,100,1))
permittivity += circle_mask[:,:,None]*(refractive_index**2 - 1)
grid[170:270, 100:200, 0] = fdtd.Object(permittivity=permittivity, name="object")
Run Simulation
[4]:
grid.run(total_time=500)
100%|██████████| 500/500 [00:03<00:00, 142.73it/s]
Visualize
Even though visualization of a circular object is not implemented (for now), one can clearly see the focussing.
[5]:
grid.visualize(z=0)

[6]:
grid.object.inverse_permittivity.min()
[6]:
0.34602076124567477
04. Performance Profiling
We can profile the performance with a 3D FDTD simulation:
Imports
[1]:
import matplotlib.pyplot as plt
from line_profiler import LineProfiler
import fdtd
import fdtd.backend as bd
Set Backend
Let’s profile the impact of the backend. These are the possible backends:
numpy
(defaults to float64 arrays)torch
(defaults to float64 tensors)torch.float32
torch.float64
torch.cuda
(defaults to float64 tensors)torch.cuda.float32
torch.cuda.float64
[2]:
fdtd.set_backend("numpy")
In general, the numpy
backend is preferred for standard CPU calculations with "float64"
precision as it is slightly faster than torch
backend on CPU. However, a significant performance improvement can be obtained by choosing torch.cuda
on large enough grids.
Note that, in FDTD, float64
precision is generally preferred over float32
to ensure numerical stability and prevent numerical dispersion. If this is of no concern to you, you can opt for float32
precision, which especially on a GPU might yield a significant performance boost.
Constants
[3]:
WAVELENGTH = 1550e-9
SPEED_LIGHT: float = 299_792_458.0 # [m/s] speed of light
Setup Simulation
create FDTD Grid
[4]:
N = 100
grid = fdtd.Grid(
(N, N, N),
grid_spacing=0.05 * WAVELENGTH,
permittivity=1.0,
permeability=1.0,
)
add boundaries
[5]:
# x boundaries
grid[0:10, :, :] = fdtd.PML(name="pml_xlow")
grid[-10:, :, :] = fdtd.PML(name="pml_xhigh")
# y boundaries
grid[:, 0:10, :] = fdtd.PML(name="pml_ylow")
grid[:, -10:, :] = fdtd.PML(name="pml_yhigh")
# z boundaries
grid[:, :, 0:10] = fdtd.PML(name="pml_zlow")
grid[:, :, -10:] = fdtd.PML(name="pml_zhigh")
add sources
[6]:
grid[10+N//10:10+N//10, :, :] = fdtd.PlaneSource(
period=WAVELENGTH / SPEED_LIGHT, name="source"
)
add objects
[7]:
grid[10+N//5:4*N//5-10, 10+N//5:4*N//5-10, 10+N//5:4*N//5-10] = fdtd.Object(permittivity=2.5, name="center_object")
grid summary
[8]:
print(grid)
Grid(shape=(100,100,100), grid_spacing=7.75e-08, courant_number=0.57)
sources:
PlaneSource(period=35, amplitude=1.0, phase_shift=0.0, name='source', polarization='z')
@ x=[20, ... , 21], y=[0, ... , 100], z=[0, ... , 100]
boundaries:
PML(name='pml_xlow')
@ x=0:10, y=:, z=:
PML(name='pml_xhigh')
@ x=-10:, y=:, z=:
PML(name='pml_ylow')
@ x=:, y=0:10, z=:
PML(name='pml_yhigh')
@ x=:, y=-10:, z=:
PML(name='pml_zlow')
@ x=:, y=:, z=0:10
PML(name='pml_zhigh')
@ x=:, y=:, z=-10:
objects:
Object(name='center_object')
@ x=30:70, y=30:70, z=30:70
Setup LineProfiler
create and enable profiler
[9]:
profiler = LineProfiler()
profiler.add_function(grid.update_E)
profiler.enable()
Run Simulation
run simulation
[10]:
grid.run(50, progress_bar=True)
100%|██████████| 50/50 [00:13<00:00, 3.85it/s]
Profiler Results
print profiler summary
[11]:
profiler.print_stats()
Timer unit: 1e-09 s
Total time: 6.46261 s
File: /home/docs/checkouts/readthedocs.org/user_builds/fdtd/checkouts/latest/docs/examples/fdtd/grid.py
Function: update_E at line 275
Line # Hits Time Per Hit % Time Line Contents
==============================================================
275 def update_E(self):
276 """update the electric field by using the curl of the magnetic field"""
277
278 # update boundaries: step 1
279 300 484243.0 1614.1 0.0 for boundary in self.boundaries:
280 300 2407036822.0 8023456.1 37.2 boundary.update_phi_E()
281
282 50 2753370607.0 55067412.1 42.6 curl = curl_H(self.H)
283 50 778202373.0 15564047.5 12.0 self.E += self.courant_number * self.inverse_permittivity * curl
284
285 # update objects
286 50 202203.0 4044.1 0.0 for obj in self.objects:
287 50 73191177.0 1463823.5 1.1 obj.update_E(curl)
288
289 # update boundaries: step 2
290 300 336969.0 1123.2 0.0 for boundary in self.boundaries:
291 300 446512390.0 1488374.6 6.9 boundary.update_E()
292
293 # add sources to grid:
294 50 85797.0 1715.9 0.0 for src in self.sources:
295 50 3141806.0 62836.1 0.0 src.update_E()
296
297 # detect electric field
298 50 46956.0 939.1 0.0 for det in self.detectors:
299 det.detect_E()
Visualization
[12]:
plt.figure()
grid.visualize(z=N//2)

05. Lenses and analysing lensing actions
submitted by substancia, adapted by flaport
Imports
[1]:
import os
import fdtd
import numpy as np
import matplotlib.pyplot as plt
Grid
[2]:
grid = fdtd.Grid(shape=(260, 15.5e-6, 1), grid_spacing=77.5e-9)
# x boundaries
grid[0:10, :, :] = fdtd.PML(name="pml_xlow")
grid[-10:, :, :] = fdtd.PML(name="pml_xhigh")
# y boundaries
grid[:, 0:10, :] = fdtd.PML(name="pml_ylow")
grid[:, -10:, :] = fdtd.PML(name="pml_yhigh")
simfolder = grid.save_simulation("Lenses") # initializing environment to save simulation data
print(simfolder)
/home/docs/checkouts/readthedocs.org/user_builds/fdtd/checkouts/latest/docs/examples/fdtd_output/fdtd_output_2023-5-23-13-17-1 (Lenses)
Objects
defining a biconvex lens
[3]:
x, y = np.arange(-200, 200, 1), np.arange(190, 200, 1)
X, Y = np.meshgrid(x, y)
lens_mask = X ** 2 + Y ** 2 <= 40000
for j, col in enumerate(lens_mask.T):
for i, val in enumerate(np.flip(col)):
if val:
grid[30 + i : 50 - i, j - 100 : j - 99, 0] = fdtd.Object(permittivity=1.5 ** 2, name=str(i) + "," + str(j))
break
Source
using a continuous source (not a pulse)
[4]:
grid[15, 50:150, 0] = fdtd.LineSource(period=1550e-9 / (3e8), name="source")
Detectors
using a BlockDetector
[5]:
grid[80:200, 80:120, 0] = fdtd.BlockDetector(name="detector")
Saving grid geometry for future reference
[6]:
with open(os.path.join(simfolder, "grid.txt"), "w") as f:
f.write(str(grid))
wavelength = 3e8/grid.source.frequency
wavelengthUnits = wavelength/grid.grid_spacing
GD = np.array([grid.x, grid.y, grid.z])
gridRange = [np.arange(x/grid.grid_spacing) for x in GD]
objectRange = np.array([[gridRange[0][x.x], gridRange[1][x.y], gridRange[2][x.z]] for x in grid.objects], dtype=object).T
f.write("\n\nGrid details (in wavelength scale):")
f.write("\n\tGrid dimensions: ")
f.write(str(GD/wavelength))
f.write("\n\tSource dimensions: ")
f.write(str(np.array([grid.source.x[-1] - grid.source.x[0] + 1, grid.source.y[-1] - grid.source.y[0] + 1, grid.source.z[-1] - grid.source.z[0] + 1])/wavelengthUnits))
f.write("\n\tObject dimensions: ")
f.write(str([(max(map(max, x)) - min(map(min, x)) + 1)/wavelengthUnits for x in objectRange]))
Simulation
[7]:
from IPython.display import clear_output # only necessary in jupyter notebooks
for i in range(400):
grid.step() # running simulation 1 timestep a time and animating
if i % 10 == 0:
# saving frames during visualization
grid.visualize(z=0, animate=True, index=i, save=True, folder=simfolder)
plt.title(f"{i:3.0f}")
clear_output(wait=True) # only necessary in jupyter notebooks
grid.save_data() # saving detector readings

We can generate a video with ffmpeg:
[8]:
try:
video_path = grid.generate_video(delete_frames=False) # rendering video from saved frames
except:
video_path = ""
print("ffmpeg not installed?")
ffmpeg not installed?
[9]:
if video_path:
from IPython.display import Video
display(Video(video_path, embed=True))
Analyse
analysing data stored by above simulation by plotting a 2D decibel map
[10]:
df = np.load(os.path.join(simfolder, "detector_readings.npz"))
fdtd.dB_map_2D(df["detector (E)"])
100%|██████████| 121/121 [00:01<00:00, 86.76it/s]
Peak at: [[[45, 20]]]

06. GRIN medium and analysing refraction
submitted by substancia, adapted by flaport
Imports
[1]:
import os
import fdtd
import numpy as np
import matplotlib.pyplot as plt
Grid
[2]:
grid = fdtd.Grid(shape=(9.3e-6, 15.5e-6, 1), grid_spacing=77.5e-9)
# x boundaries
grid[0:10, :, :] = fdtd.PML(name="pml_xlow")
grid[-10:, :, :] = fdtd.PML(name="pml_xhigh")
# y boundaries
grid[:, 0:10, :] = fdtd.PML(name="pml_ylow")
grid[:, -10:, :] = fdtd.PML(name="pml_yhigh")
simfolder = grid.save_simulation("GRIN") # initializing environment to save simulation data
print(simfolder)
/home/docs/checkouts/readthedocs.org/user_builds/fdtd/checkouts/latest/docs/examples/fdtd_output/fdtd_output_2023-5-23-13-18-11 (GRIN)
Objects
defining a graded refractive index slab, with homogenous slab extensions outwards from both ends
[3]:
n0, theta, t = 1, 30, 0.5
for i in range(50):
x = i * 0.08
epsilon = n0 + x * np.sin(np.radians(theta)) / t
epsilon = epsilon ** 0.5
grid[
5.1e-6:5.6e-6, (5 + i * 0.08) * 1e-6 : (5.08 + i * 0.08) * 1e-6, 0
] = fdtd.Object(permittivity=epsilon, name="object" + str(i))
# homogenous slab extensions
grid[5.1e-6:5.6e-6, 0.775e-6:5e-6, 0] = fdtd.Object(
permittivity=n0 ** 2, name="objectLeft"
)
grid[5.1e-6:5.6e-6, 9e-6 : (15.5 - 0.775) * 1e-6, 0] = fdtd.Object(
permittivity=epsilon, name="objectRight"
)
Source
using a pulse (hanning window pulse)
[4]:
grid[3.1e-6, 1.5e-6:14e-6, 0] = fdtd.LineSource(period=1550e-9 / (3e8), name="source", pulse=True, cycle=3, hanning_dt=4e-15)
Detectors
using a linear array of LineDetector
[5]:
for i in range(-4, 8):
grid[5.8e-6, 84 + 4 * i : 86 + 4 * i, 0] = fdtd.LineDetector(name="detector" + str(i))
Saving grid geometry
[6]:
with open(os.path.join("./fdtd_output", grid.folder, "grid.txt"), "w") as f:
f.write(str(grid))
wavelength = 3e8/grid.source.frequency
wavelengthUnits = wavelength/grid.grid_spacing
GD = np.array([grid.x, grid.y, grid.z])
gridRange = [np.arange(x/grid.grid_spacing) for x in GD]
objectRange = np.array([[gridRange[0][x.x], gridRange[1][x.y], gridRange[2][x.z]] for x in grid.objects], dtype=object).T
f.write("\n\nGrid details (in wavelength scale):")
f.write("\n\tGrid dimensions: ")
f.write(str(GD/wavelength))
f.write("\n\tSource dimensions: ")
f.write(str(np.array([grid.source.x[-1] - grid.source.x[0] + 1, grid.source.y[-1] - grid.source.y[0] + 1, grid.source.z[-1] - grid.source.z[0] + 1])/wavelengthUnits))
f.write("\n\tObject dimensions: ")
f.write(str([(max(map(max, x)) - min(map(min, x)) + 1)/wavelengthUnits for x in objectRange]))
Simulation
[7]:
from IPython.display import clear_output # only necessary in jupyter notebooks
for i in range(100):
grid.step() # running simulation 1 timestep a time and animating
if i % 5 == 0:
# saving frames during visualization
grid.visualize(z=0, animate=True, index=i, save=True, folder=simfolder)
plt.title(f"{i:3.0f}")
clear_output(wait=True) # only necessary in jupyter notebooks
grid.save_data() # saving detector readings

We can generate a video with ffmpeg:
[8]:
try:
video_path = grid.generate_video(delete_frames=False) # rendering video from saved frames
except:
video_path = ""
print("ffmpeg not installed?")
ffmpeg not installed?
[9]:
if video_path:
from IPython.display import Video
display(Video(video_path, embed=True))
Analyse
analysing data stored by above simulation to find intensity profile and time-of-arrival plot
[10]:
dic = np.load(os.path.join(simfolder, "detector_readings.npz"))
import warnings; warnings.filterwarnings("ignore") # TODO: fix plot_detection to prevent warnings
fdtd.plot_detection(dic)




fdtd package
backend module
Selects the backend for the fdtd-package.
The fdtd library allows to choose a backend. The numpy
backend is the
default one, but there are also several additional PyTorch backends:
numpy
(defaults to float64 arrays)
torch
(defaults to float64 tensors)
torch.float32
torch.float64
torch.cuda
(defaults to float64 tensors)
torch.cuda.float32
torch.cuda.float64
For example, this is how to choose the “torch” backend:
fdtd.set_backend("torch")
In general, the numpy
backend is preferred for standard CPU calculations
with “float64” precision. In general, float64
precision is always
preferred over float32
for FDTD simulations, however, float32
might
give a significant performance boost.
The cuda
backends are only available for computers with a GPU.
- class fdtd.backend.NumpyBackend[source]
Bases:
Backend
Numpy Backend
- arange([start, ]stop, [step, ]dtype=None, *, like=None)
create a range of values
- array()
create an array from an array-like sequence
- asarray(a, dtype=None, order=None, *, like=None)
Convert the input to an array.
- Parameters:
a (array_like) – Input data, in any form that can be converted to an array. This includes lists, lists of tuples, tuples, tuples of tuples, tuples of lists and ndarrays.
dtype (data-type, optional) – By default, the data-type is inferred from the input data.
order ({'C', 'F', 'A', 'K'}, optional) – Memory layout. ‘A’ and ‘K’ depend on the order of input array a. ‘C’ row-major (C-style), ‘F’ column-major (Fortran-style) memory representation. ‘A’ (any) means ‘F’ if a is Fortran contiguous, ‘C’ otherwise ‘K’ (keep) preserve input order Defaults to ‘K’.
like (array_like, optional) –
Reference object to allow the creation of arrays which are not NumPy arrays. If an array-like passed in as
like
supports the__array_function__
protocol, the result will be defined by it. In this case, it ensures the creation of an array object compatible with that passed in via this argument.New in version 1.20.0.
- Returns:
out – Array interpretation of a. No copy is performed if the input is already an ndarray with matching dtype and order. If a is a subclass of ndarray, a base class ndarray is returned.
- Return type:
ndarray
See also
asanyarray
Similar function which passes through subclasses.
ascontiguousarray
Convert input to a contiguous array.
asfarray
Convert input to a floating point ndarray.
asfortranarray
Convert input to an ndarray with column-major memory order.
asarray_chkfinite
Similar function which checks input for NaNs and Infs.
fromiter
Create an array from an iterator.
fromfunction
Construct an array by executing a function on grid positions.
Examples
Convert a list into an array:
>>> a = [1, 2] >>> np.asarray(a) array([1, 2])
Existing arrays are not copied:
>>> a = np.array([1, 2]) >>> np.asarray(a) is a True
If dtype is set, array is copied only if dtype does not match:
>>> a = np.array([1, 2], dtype=np.float32) >>> np.asarray(a, dtype=np.float32) is a True >>> np.asarray(a, dtype=np.float64) is a False
Contrary to asanyarray, ndarray subclasses are not passed through:
>>> issubclass(np.recarray, np.ndarray) True >>> a = np.array([(1.0, 2), (3.0, 4)], dtype='f4,i4').view(np.recarray) >>> np.asarray(a) is a False >>> np.asanyarray(a) is a True
- static broadcast_arrays(*args, subok=False)
broadcast arrays
- static broadcast_to(array, shape, subok=False)
broadcast array into shape
- complex
complex type for array
alias of
complex128
- cos = <ufunc 'cos'>
cosine of all elements in array
- divide = <ufunc 'divide'>
- exp = <ufunc 'exp'>
exponential of all elements in array
- static fft(a, n=None, axis=-1, norm=None)
Compute the one-dimensional discrete Fourier Transform.
This function computes the one-dimensional n-point discrete Fourier Transform (DFT) with the efficient Fast Fourier Transform (FFT) algorithm [CT].
- Parameters:
a (array_like) – Input array, can be complex.
n (int, optional) – Length of the transformed axis of the output. If n is smaller than the length of the input, the input is cropped. If it is larger, the input is padded with zeros. If n is not given, the length of the input along the axis specified by axis is used.
axis (int, optional) – Axis over which to compute the FFT. If not given, the last axis is used.
norm ({"backward", "ortho", "forward"}, optional) –
New in version 1.10.0.
Normalization mode (see numpy.fft). Default is “backward”. Indicates which direction of the forward/backward pair of transforms is scaled and with what normalization factor.
New in version 1.20.0: The “backward”, “forward” values were added.
- Returns:
out – The truncated or zero-padded input, transformed along the axis indicated by axis, or the last one if axis is not specified.
- Return type:
complex ndarray
- Raises:
IndexError – If axis is not a valid axis of a.
See also
numpy.fft
for definition of the DFT and conventions used.
ifft
The inverse of fft.
fft2
The two-dimensional FFT.
fftn
The n-dimensional FFT.
rfftn
The n-dimensional FFT of real input.
fftfreq
Frequency bins for given FFT parameters.
Notes
FFT (Fast Fourier Transform) refers to a way the discrete Fourier Transform (DFT) can be calculated efficiently, by using symmetries in the calculated terms. The symmetry is highest when n is a power of 2, and the transform is therefore most efficient for these sizes.
The DFT is defined, with the conventions used in this implementation, in the documentation for the numpy.fft module.
References
[CT]Cooley, James W., and John W. Tukey, 1965, “An algorithm for the machine calculation of complex Fourier series,” Math. Comput. 19: 297-301.
Examples
>>> np.fft.fft(np.exp(2j * np.pi * np.arange(8) / 8)) array([-2.33486982e-16+1.14423775e-17j, 8.00000000e+00-1.25557246e-15j, 2.33486982e-16+2.33486982e-16j, 0.00000000e+00+1.22464680e-16j, -1.14423775e-17+2.33486982e-16j, 0.00000000e+00+5.20784380e-16j, 1.14423775e-17+1.14423775e-17j, 0.00000000e+00+1.22464680e-16j])
In this example, real input has an FFT which is Hermitian, i.e., symmetric in the real part and anti-symmetric in the imaginary part, as described in the numpy.fft documentation:
>>> import matplotlib.pyplot as plt >>> t = np.arange(256) >>> sp = np.fft.fft(np.sin(t)) >>> freq = np.fft.fftfreq(t.shape[-1]) >>> plt.plot(freq, sp.real, freq, sp.imag) [<matplotlib.lines.Line2D object at 0x...>, <matplotlib.lines.Line2D object at 0x...>] >>> plt.show()
- static fftfreq(n, d=1.0)
Return the Discrete Fourier Transform sample frequencies.
The returned float array f contains the frequency bin centers in cycles per unit of the sample spacing (with zero at the start). For instance, if the sample spacing is in seconds, then the frequency unit is cycles/second.
Given a window length n and a sample spacing d:
f = [0, 1, ..., n/2-1, -n/2, ..., -1] / (d*n) if n is even f = [0, 1, ..., (n-1)/2, -(n-1)/2, ..., -1] / (d*n) if n is odd
- Parameters:
n (int) – Window length.
d (scalar, optional) – Sample spacing (inverse of the sampling rate). Defaults to 1.
- Returns:
f – Array of length n containing the sample frequencies.
- Return type:
ndarray
Examples
>>> signal = np.array([-2, 8, 6, 4, 1, 0, 3, 5], dtype=float) >>> fourier = np.fft.fft(signal) >>> n = signal.size >>> timestep = 0.1 >>> freq = np.fft.fftfreq(n, d=timestep) >>> freq array([ 0. , 1.25, 2.5 , ..., -3.75, -2.5 , -1.25])
- float
floating type for array
alias of
float64
- int
integer type for array
alias of
int64
- linspace(stop, num=50, endpoint=True, retstep=False, dtype=None, axis=0)
create a linearly spaced array between two points
- static max(a, axis=None, out=None, keepdims=<no value>, initial=<no value>, where=<no value>)
max element in array
- numpy()
convert the array to numpy array
- ones(dtype=None, order='C', *, like=None)
create an array filled with ones
- static pad(array, pad_width, mode='constant', **kwargs)
Pad an array.
- Parameters:
array (array_like of rank N) – The array to pad.
pad_width ({sequence, array_like, int}) – Number of values padded to the edges of each axis.
((before_1, after_1), ... (before_N, after_N))
unique pad widths for each axis.(before, after)
or((before, after),)
yields same before and after pad for each axis.(pad,)
orint
is a shortcut for before = after = pad width for all axes.mode (str or function, optional) –
One of the following string values or a user supplied function.
- ’constant’ (default)
Pads with a constant value.
- ’edge’
Pads with the edge values of array.
- ’linear_ramp’
Pads with the linear ramp between end_value and the array edge value.
- ’maximum’
Pads with the maximum value of all or part of the vector along each axis.
- ’mean’
Pads with the mean value of all or part of the vector along each axis.
- ’median’
Pads with the median value of all or part of the vector along each axis.
- ’minimum’
Pads with the minimum value of all or part of the vector along each axis.
- ’reflect’
Pads with the reflection of the vector mirrored on the first and last values of the vector along each axis.
- ’symmetric’
Pads with the reflection of the vector mirrored along the edge of the array.
- ’wrap’
Pads with the wrap of the vector along the axis. The first values are used to pad the end and the end values are used to pad the beginning.
- ’empty’
Pads with undefined values.
New in version 1.17.
- <function>
Padding function, see Notes.
stat_length (sequence or int, optional) –
Used in ‘maximum’, ‘mean’, ‘median’, and ‘minimum’. Number of values at edge of each axis used to calculate the statistic value.
((before_1, after_1), ... (before_N, after_N))
unique statistic lengths for each axis.(before, after)
or((before, after),)
yields same before and after statistic lengths for each axis.(stat_length,)
orint
is a shortcut forbefore = after = statistic
length for all axes.Default is
None
, to use the entire axis.constant_values (sequence or scalar, optional) –
Used in ‘constant’. The values to set the padded values for each axis.
((before_1, after_1), ... (before_N, after_N))
unique pad constants for each axis.(before, after)
or((before, after),)
yields same before and after constants for each axis.(constant,)
orconstant
is a shortcut forbefore = after = constant
for all axes.Default is 0.
end_values (sequence or scalar, optional) –
Used in ‘linear_ramp’. The values used for the ending value of the linear_ramp and that will form the edge of the padded array.
((before_1, after_1), ... (before_N, after_N))
unique end values for each axis.(before, after)
or((before, after),)
yields same before and after end values for each axis.(constant,)
orconstant
is a shortcut forbefore = after = constant
for all axes.Default is 0.
reflect_type ({'even', 'odd'}, optional) – Used in ‘reflect’, and ‘symmetric’. The ‘even’ style is the default with an unaltered reflection around the edge value. For the ‘odd’ style, the extended part of the array is created by subtracting the reflected values from two times the edge value.
- Returns:
pad – Padded array of rank equal to array with shape increased according to pad_width.
- Return type:
ndarray
Notes
New in version 1.7.0.
For an array with rank greater than 1, some of the padding of later axes is calculated from padding of previous axes. This is easiest to think about with a rank 2 array where the corners of the padded array are calculated by using padded values from the first axis.
The padding function, if used, should modify a rank 1 array in-place. It has the following signature:
padding_func(vector, iaxis_pad_width, iaxis, kwargs)
where
- vectorndarray
A rank 1 array already padded with zeros. Padded values are vector[:iaxis_pad_width[0]] and vector[-iaxis_pad_width[1]:].
- iaxis_pad_widthtuple
A 2-tuple of ints, iaxis_pad_width[0] represents the number of values padded at the beginning of vector where iaxis_pad_width[1] represents the number of values padded at the end of vector.
- iaxisint
The axis currently being calculated.
- kwargsdict
Any keyword arguments the function requires.
Examples
>>> a = [1, 2, 3, 4, 5] >>> np.pad(a, (2, 3), 'constant', constant_values=(4, 6)) array([4, 4, 1, ..., 6, 6, 6])
>>> np.pad(a, (2, 3), 'edge') array([1, 1, 1, ..., 5, 5, 5])
>>> np.pad(a, (2, 3), 'linear_ramp', end_values=(5, -4)) array([ 5, 3, 1, 2, 3, 4, 5, 2, -1, -4])
>>> np.pad(a, (2,), 'maximum') array([5, 5, 1, 2, 3, 4, 5, 5, 5])
>>> np.pad(a, (2,), 'mean') array([3, 3, 1, 2, 3, 4, 5, 3, 3])
>>> np.pad(a, (2,), 'median') array([3, 3, 1, 2, 3, 4, 5, 3, 3])
>>> a = [[1, 2], [3, 4]] >>> np.pad(a, ((3, 2), (2, 3)), 'minimum') array([[1, 1, 1, 2, 1, 1, 1], [1, 1, 1, 2, 1, 1, 1], [1, 1, 1, 2, 1, 1, 1], [1, 1, 1, 2, 1, 1, 1], [3, 3, 3, 4, 3, 3, 3], [1, 1, 1, 2, 1, 1, 1], [1, 1, 1, 2, 1, 1, 1]])
>>> a = [1, 2, 3, 4, 5] >>> np.pad(a, (2, 3), 'reflect') array([3, 2, 1, 2, 3, 4, 5, 4, 3, 2])
>>> np.pad(a, (2, 3), 'reflect', reflect_type='odd') array([-1, 0, 1, 2, 3, 4, 5, 6, 7, 8])
>>> np.pad(a, (2, 3), 'symmetric') array([2, 1, 1, 2, 3, 4, 5, 5, 4, 3])
>>> np.pad(a, (2, 3), 'symmetric', reflect_type='odd') array([0, 1, 1, 2, 3, 4, 5, 5, 6, 7])
>>> np.pad(a, (2, 3), 'wrap') array([4, 5, 1, 2, 3, 4, 5, 1, 2, 3])
>>> def pad_with(vector, pad_width, iaxis, kwargs): ... pad_value = kwargs.get('padder', 10) ... vector[:pad_width[0]] = pad_value ... vector[-pad_width[1]:] = pad_value >>> a = np.arange(6) >>> a = a.reshape((2, 3)) >>> np.pad(a, 2, pad_with) array([[10, 10, 10, 10, 10, 10, 10], [10, 10, 10, 10, 10, 10, 10], [10, 10, 0, 1, 2, 10, 10], [10, 10, 3, 4, 5, 10, 10], [10, 10, 10, 10, 10, 10, 10], [10, 10, 10, 10, 10, 10, 10]]) >>> np.pad(a, 2, pad_with, padder=100) array([[100, 100, 100, 100, 100, 100, 100], [100, 100, 100, 100, 100, 100, 100], [100, 100, 0, 1, 2, 100, 100], [100, 100, 3, 4, 5, 100, 100], [100, 100, 100, 100, 100, 100, 100], [100, 100, 100, 100, 100, 100, 100]])
- static reshape(a, newshape, order='C')
reshape array into given shape
- sin = <ufunc 'sin'>
sine of all elements in array
- static squeeze(a, axis=None)
remove dim-1 dimensions
- static stack(arrays, axis=0, out=None, *, dtype=None, casting='same_kind')
stack multiple arrays
- static sum(a, axis=None, dtype=None, out=None, keepdims=<no value>, initial=<no value>, where=<no value>)
sum elements in array
- static transpose(a, axes=None)
transpose array by flipping two dimensions
- zeros(shape, dtype=float, order='C', *, like=None)
create an array filled with zeros
- static zeros_like(a, dtype=None, order='K', subok=True, shape=None)
create an array filled with zeros
- fdtd.backend.set_backend(name: str)[source]
Set the backend for the FDTD simulations
This function monkeypatches the backend object by changing its class. This way, all methods of the backend object will be replaced.
- Parameters:
name – name of the backend. Allowed backend names: -
numpy
(defaults to float64 arrays) -numpy.float16
-numpy.float32
-numpy.float64
-numpy.float128
-torch
(defaults to float64 tensors) -torch.float16
-torch.float32
-torch.float64
-torch.cuda
(defaults to float64 tensors) -torch.cuda.float16
-torch.cuda.float32
-torch.cuda.float64
boundaries module
Boundaries for the FDTD Grid.
Available Boundaries:
PeriodicBoundary
PML
- class fdtd.boundaries.Boundary(name: str | None = None)[source]
Bases:
object
an FDTD Boundary [base class]
- __init__(name: str | None = None)[source]
Create a boundary
- Parameters:
name – name of the boundary
- update_E()[source]
Update electric field of the grid
Note
this method is called after the grid fields are updated
- update_H()[source]
Update magnetic field of the grid
Note
this method is called after the grid fields are updated
- fdtd.boundaries.DomainBorderPML(grid, border_cells=5)[source]
Some problem setups require a layer of PML all the way around the problem. This is a convenience function to add such a layer to an existing grid. Caution: Alters grid in-place.
- class fdtd.boundaries.PML(a: float = 1e-08, name: str | None = None)[source]
Bases:
Boundary
A perfectly matched layer (PML)
a PML is an impedence-matched area at the boundary of the grid for which all fields incident perpendicular to the area are absorbed without reflection.
Note
Registering a PML to the grid will monkeypatch the PML to become one of its subclasses:
_PMLXlow
,_PMLYlow
or_PMLZlow
,_PMLXhigh
,_PMLYhigh
,_PMLZhigh
depending on the position in the grid.- __init__(a: float = 1e-08, name: str | None = None)[source]
Perfectly Matched Layer
- Parameters:
a – stability parameter
name – name of the PML
- update_E()[source]
Update electric field of the grid
Note
this method is called after the electric field is updated
- update_H()[source]
Update magnetic field of the grid
Note
this method is called after the magnetic field is updated
- class fdtd.boundaries.PeriodicBoundary(name: str | None = None)[source]
Bases:
Boundary
An FDTD Periodic Boundary
Note
Registering a periodic boundary to the grid will change the periodic boundary in one of its subclasses:
_PeriodicBoundaryX
,_PeriodicBoundaryY
or_PeriodicBoundaryY
, depending on the position in the grid.
detectors module
Detectors for the FDTD Grid.
Available Detectors:
LineDetector
- class fdtd.detectors.BlockDetector(name=None)[source]
Bases:
object
A detector along a block in the FDTD grid
- class fdtd.detectors.CurrentDetector(name=None)[source]
Bases:
object
A current detector.
- single_point_current(px, py, pz)[source]
Only Z-polarized for now. Can probably do a cross product to get arbitrary polarizations
^
X—->
TODO: FIXME: IMPORTANT: material magnetic permeability? find test cases!
Implements the first correction from [Fang 1994] (two cells are spatially averaged to account for Yee cell half-step inaccuracies), but not the second one (minor loss of accuracy).
Jiayuan Fang, Danwei Xue. Precautions in the calculation of impedance in FDTD computations. Proceedings of IEEE Antennas and Propagation Society International Symposium and URSI National Radio Science Meeting, vol. 3, 1994, p. 1814–7 vol.3. https://doi.org/10.1109/APS.1994.408185.
Luebbers RJ, Langdon HS. A simple feed model that reduces time steps needed for FDTD antenna and microstrip calculations. IEEE Trans Antennas Propagat 1996;44:1000–5. https://doi.org/10.1109/8.504308.
grid module
The FDTD Grid
The grid is the core of the FDTD Library. It is where everything comes together and where the biggest part of the calculations are done.
- class fdtd.grid.Grid(shape: Tuple[Number, Number, Number], grid_spacing: float = 1.55e-07, permittivity: float = 1.0, permeability: float = 1.0, courant_number: float | None = None)[source]
Bases:
object
The FDTD Grid
The grid is the core of the FDTD Library. It is where everything comes together and where the biggest part of the calculations are done.
- __init__(shape: Tuple[Number, Number, Number], grid_spacing: float = 1.55e-07, permittivity: float = 1.0, permeability: float = 1.0, courant_number: float | None = None)[source]
- Parameters:
shape – shape of the FDTD grid.
grid_spacing – distance between the grid cells.
permittivity – the relative permittivity of the background.
permeability – the relative permeability of the background.
courant_number – the courant number of the FDTD simulation. Defaults to the inverse of the square root of the number of dimensions > 1 (optimal value). The timestep of the simulation will be derived from this number using the CFL-condition.
- generate_video(delete_frames=False)[source]
Compiles frames into a video
These framed should be saved through
fdtd.Grid.visualize(save=True)
while havingfdtd.Grid.save_simulation()
enabled.- Parameters:
delete_frames (optional, bool) – delete stored frames after conversion to video.
- Returns:
the filename of the generated video.
Note
this function requires
ffmpeg
to be available in your path.
- run(total_time: Number, progress_bar: bool = True)[source]
run an FDTD simulation.
- Parameters:
total_time – the total time for the simulation to run.
progress_bar – choose to show a progress bar during simulation
- save_data()[source]
Saves readings from all detectors in the grid into a numpy zip file. Each detector is stored in separate arrays. Electric and magnetic field field readings of each detector are also stored separately with suffix ” (E)” and ” (H)” (Example: [‘detector0 (E)’, ‘detector0 (H)’]). Therefore, the numpy zip file contains arrays twice the number of detectors. REQUIRES ‘fdtd.Grid.save_simulation()’ to be run before this function.
Parameters: None
- save_simulation(sim_name=None)[source]
Creates a folder and initializes environment to store simulation or related details. saveSimulation() needs to be run before running any function that stores data (generate_video(), save_data()).
- Parameters:-
(optional) sim_name (string): Preferred name for simulation
- property shape: Tuple[int, int, int]
get the shape of the FDTD grid
- step()[source]
do a single FDTD step by first updating the electric field and then updating the magnetic field
- property time_passed: float
get the total time passed
- visualize(x=None, y=None, z=None, cmap='Blues', pbcolor='C3', pmlcolor=(0, 0, 0, 0.1), objcolor=(1, 0, 0, 0.1), srccolor='C0', detcolor='C2', norm='linear', show=False, animate=False, index=None, save=False, folder=None)
visualize a projection of the grid and the optical energy inside the grid
- Parameters:
x – the x-value to make the yz-projection (leave None if using different projection)
y – the y-value to make the zx-projection (leave None if using different projection)
z – the z-value to make the xy-projection (leave None if using different projection)
cmap – the colormap to visualize the energy in the grid
pbcolor – the color to visualize the periodic boundaries
pmlcolor – the color to visualize the PML
objcolor – the color to visualize the objects in the grid
srccolor – the color to visualize the sources in the grid
detcolor – the color to visualize the detectors in the grid
norm – how to normalize the grid_energy color map (‘linear’ or ‘log’).
show – call pyplot.show() at the end of the function
animate – see frame by frame state of grid during simulation
index – index for each frame of animation (typically a loop variable is passed)
save – save frames in a folder
folder – path to folder to save frames
- property x: int
get the number of grid cells in the x-direction
- property y: int
get the number of grid cells in the y-direction
- property z: int
get the number of grid cells in the y-direction
- fdtd.grid.curl_E(E: ndarray) ndarray [source]
Transforms an E-type field into an H-type field by performing a curl operation
- Parameters:
E – Electric field to take the curl of (E-type field located on the edges of the grid cell [integer gridpoints])
- Returns:
The curl of E (H-type field located on the faces of the grid [half-integer grid points])
- fdtd.grid.curl_H(H: ndarray) ndarray [source]
Transforms an H-type field into an E-type field by performing a curl operation
- Parameters:
H – Magnetic field to take the curl of (H-type field located on half-integer grid points)
- Returns:
The curl of H (E-type field located on the edges of the grid [integer grid points])
objects module
The objects to place in the grid.
Objects define all the regions in the grid with a modified update equation, such as for example regions with anisotropic permittivity etc.
- Available Objects:
Object
AnisotropicObject
- class fdtd.objects.AbsorbingObject(permittivity: ndarray, conductivity: ndarray, name: str | None = None)[source]
Bases:
Object
An absorbing object takes conductivity into account
- __init__(permittivity: ndarray, conductivity: ndarray, name: str | None = None)[source]
- Parameters:
permittivity – permittivity tensor
conductivity – conductivity tensor (will introduce the loss)
name – name of the object (will become available as attribute to the grid)
- class fdtd.objects.AnisotropicObject(permittivity: ndarray, name: str | None = None)[source]
Bases:
Object
An object with anisotropic permittivity tensor
- class fdtd.objects.Object(permittivity: ndarray, name: str | None = None)[source]
Bases:
object
An object to place in the grid
- __init__(permittivity: ndarray, name: str | None = None)[source]
- Parameters:
permittivity – permittivity tensor
name – name of the object (will become available as attribute to the grid)
sources module
Sources are objects that inject the fields into the grid.
Available sources:
PointSource
LineSource
- class fdtd.sources.LineSource(period: Number = 15, amplitude: float = 1.0, phase_shift: float = 0.0, name: str | None = None, pulse: bool = False, cycle: int = 5, hanning_dt: float = 10.0)[source]
Bases:
object
A source along a line in the FDTD grid
- __init__(period: Number = 15, amplitude: float = 1.0, phase_shift: float = 0.0, name: str | None = None, pulse: bool = False, cycle: int = 5, hanning_dt: float = 10.0)[source]
Create a LineSource with a gaussian profile
- Parameters:
period – The period of the source. The period can be specified as integer [timesteps] or as float [seconds]
amplitude – The amplitude of the source in simulation units
phase_shift – The phase offset of the source.
pulse – Set True to use a Hanning window pulse instead of continuous wavefunction.
cycle – cycles for Hanning window pulse.
hanning_dt – timestep used for Hanning window pulse width (optional).
- class fdtd.sources.PlaneSource(period: Number = 15, amplitude: float = 1.0, phase_shift: float = 0.0, name: str | None = None, polarization: str = 'z')[source]
Bases:
object
A source along a plane in the FDTD grid
- __init__(period: Number = 15, amplitude: float = 1.0, phase_shift: float = 0.0, name: str | None = None, polarization: str = 'z')[source]
Create a PlaneSource.
- Parameters:
period – The period of the source. The period can be specified as integer [timesteps] or as float [seconds]
amplitude – The amplitude of the source in simulation units
phase_shift – The phase offset of the source.
polarization – Axis of E-field polarization (‘x’,’y’,or ‘z’)
- class fdtd.sources.PointSource(period: Number = 15, amplitude: float = 1.0, phase_shift: float = 0.0, name: str | None = None, pulse: bool = False, cycle: int = 5, hanning_dt: float = 10.0)[source]
Bases:
object
A source placed at a single point (grid cell) in the grid
- __init__(period: Number = 15, amplitude: float = 1.0, phase_shift: float = 0.0, name: str | None = None, pulse: bool = False, cycle: int = 5, hanning_dt: float = 10.0)[source]
Create a LineSource with a gaussian profile
- Parameters:
period – The period of the source. The period can be specified as integer [timesteps] or as float [seconds]
amplitude – The electric field amplitude in simulation units
phase_shift – The phase offset of the source.
name – name of the source.
pulse – Set True to use a Hanning window pulse instead of continuous wavefunction.
cycle – cycles for Hanning window pulse.
hanning_dt – timestep used for Hanning window pulse width (optional).
- class fdtd.sources.SoftArbitraryPointSource(waveform_array: ndarray, name: str | None = None, impedance: float = 0.0)[source]
Bases:
object
A source placed at a single point (grid cell) in the grid. This source is special: it’s both a source and a detector.
Unlike the other sources, the input is a voltage, not an electric field. (really? why? should we convert back and forth?)
For electrical measurements I’ve only needed a single-index source, so I don’t know how the volume/line sources above work. We want the FFT function to operate over any detector. Maybe all sources should take an arbitary waveform argument?
Each index in the waveform array represents 1 value at a timestep.
There are many different geometries of “equivalent sources”. The detector/source paradigm used in /fdtd might perhaps not correspond to this in an ideal fashion.
It’s not intuitively clear to me what a “soft” source would imply in the optical case, or what impedance even means for a laser.
/fdtd/ seems to have found primary use in optical circles, so the default Z should probably be 0.
“Whilst established for microwaves and electrical circuits, this concept has only very recently been observed in the optical domain, yet is not well defined or understood.”[1]
[1]: Optical impedance of metallic nano-structures, M. Mazilu and K. Dholakia https://doi.org/10.1364/OE.14.007709
[2]: http://www.gwoptics.org/learn/02_Plane_waves/01_Fabry_Perot_cavity/02_Impedance_matched.php
//-
Installation
The fdtd
-library can be installed with pip
:
pip install fdtd
Dependencies
python 3.6+
numpy
matplotlib
tqdm
pytorch (optional)
Quick intro
The fdtd
library is simply imported as follows:
import fdtd
Setting the backend
The fdtd
library allows to choose a backend. The "numpy"
backend is the
default one, but there are also several additional PyTorch backends:
numpy
(defaults to float64 arrays)torch
(defaults to float64 tensors)torch.float32
torch.float64
torch.cuda
(defaults to float64 tensors)torch.cuda.float32
torch.cuda.float64
For example, this is how to choose the torch
backend:
fdtd.set_backend("torch")
In general, the numpy
backend is preferred for standard CPU calculations
with float64
precision. In general, float64
precision is always
preferred over float32
for FDTD simulations, however, float32
might
give a significant performance boost.
The cuda
backends are only available for computers with a GPU.
The FDTD-grid
The FDTD grid defines the simulation region.
# signature
fdtd.Grid(
shape: Tuple[Number, Number, Number],
grid_spacing: float = 155e-9,
permittivity: float = 1.0,
permeability: float = 1.0,
courant_number: float = None,
)
A grid is defined by its shape
, which is just a 3D tuple of
Number
-types (integers or floats). If the shape is given in floats, it
denotes the width, height and length of the grid in meters. If the shape is
given in integers, it denotes the width, height and length of the grid in terms
of the grid_spacing
. Internally, these numbers will be translated to three
integers: grid.Nx
, grid.Ny
and grid.Nz
.
A grid_spacing
can be given. For stability reasons, it is recommended to
choose a grid spacing that is at least 10 times smaller than the _smallest_
wavelength in the grid. This means that for a grid containing a source with
wavelength 1550nm
and a material with refractive index of 3.1
, the
recommended minimum grid_spacing
turns out to be 50pm
For the permittivity
and permeability
floats or arrays with the
following shapes
(grid.Nx, grid.Ny, grid.Nz)
or
(grid.Nx, grid.Ny, grid.Nz, 1)
or
(grid.Nx, grid.Ny, grid.Nz, 3)
are expected. In the last case, the shape implies the possibility for different
permittivity for each of the major axes (so-called _uniaxial_ or _biaxial_
materials). Internally, these variables will be converted (for performance
reasons) to their inverses grid.inverse_permittivity
array and a
grid.inverse_permeability
array of shape (grid.Nx, grid.Ny, grid.Nz, 3)
. It
is possible to change those arrays after making the grid.
Finally, the courant_number
of the grid determines the relation between the
time_step
of the simulation and the grid_spacing
of the grid. If not given,
it is chosen to be the maximum number allowed by the Courant-Friedrichs-Lewy Condition:
1
for 1D
simulations, 1/√2
for 2D
simulations and 1/√3
for 3D
simulations (the dimensionality will be derived by the shape of the grid). For
stability reasons, it is recommended not to change this value.
grid = fdtd.Grid(
shape = (25e-6, 15e-6, 1), # 25um x 15um x 1 (grid_spacing) --> 2D FDTD
)
print(grid)
Grid(shape=(161,97,1), grid_spacing=1.55e-07, courant_number=0.70)
Objects
An other option to locally change the permittivity
or permeability
in the
grid is to add an Object
to the grid.
# signature
fdtd.Object(
permittivity: Tensorlike,
name: str = None
)
An object defines a part of the grid with modified update equations, allowing
to introduce for example absorbing materials or biaxial materials for which
mixing between the axes are present through Pockels coefficients
or many
more. In this case we’ll make an object with a different permittivity
than
the grid it is in.
Just like for the grid, the Object
expects a permittivity
to be a floats or
an array of the following possible shapes
(obj.Nx, obj.Ny, obj.Nz)
or
(obj.Nx, obj.Ny, obj.Nz, 1)
or
(obj.Nx, obj.Ny, obj.Nz, 3)
Note that the values obj.Nx
, obj.Ny
and obj.Nz
are not given to the
object constructor. They are in stead derived from its placing in the grid:
grid[11:32, 30:84, 0] = fdtd.Object(permittivity=1.7**2, name="object")
Several things happen here. First of all, the object is given the space
[11:32, 30:84, 0]
in the grid. Because it is given this space, the object’s
Nx
, Ny
and Nz
are automatically set. Furthermore, by supplying a name to
the object, this name will become available in the grid:
print(grid.object)
Object(name='object')
@ x=11:32, y=30:84, z=0:1
A second object can be added to the grid:
grid[13e-6:18e-6, 5e-6:8e-6, 0] = fdtd.Object(permittivity=1.5**2)
Here, a slice with floating point numbers was chosen. These floats will be
replaced by integer Nx
, Ny
and Nz
during the registration of the object.
Since the object did not receive a name, the object won’t be available as an
attribute of the grid. However, it is still available via the grid.objects
list:
print(grid.objects)
[Object(name='object'), Object(name=None)]
This list stores all objects (i.e. of type fdtd.Object
) in the order that
they were added to the grid.
Sources
Similarly as to adding an object to the grid, an fdtd.LineSource
can also
be added:
# signature
fdtd.LineSource(
period: Number = 15, # timesteps or seconds
amplitude: float = 1.0,
phase_shift: float = 0.0,
name: str = None,
)
And also just like an fdtd.Object
, an fdtd.LineSource
size is defined by its
placement on the grid:
grid[7.5e-6:8.0e-6, 11.8e-6:13.0e-6, 0] = fdtd.LineSource(
period = 1550e-9 / (3e8), name="source"
)
However, it is important to note that in this case a LineSource
is added to
the grid, i.e. the source spans the diagonal of the cube defined by the slices.
Internally, these slices will be converted into lists to ensure this behavior:
print(grid.source)
LineSource(period=14, amplitude=1.0, phase_shift=0.0, name='source')
@ x=[48, ... , 51], y=[76, ... , 83], z=[0, ... , 0]
Note that one could also have supplied lists to index the grid in the first
place. This feature could be useful to create a LineSource
of arbitrary
shape.
Detectors
# signature
fdtd.LineDetector(
name=None
)
Adding a detector to the grid works the same as adding a source
grid[12e-6, :, 0] = fdtd.LineDetector(name="detector")
print(grid.detector)
LineDetector(name='detector')
@ x=[77, ... , 77], y=[0, ... , 96], z=[0, ... , 0]
Boundaries
# signature
fdtd.PML(
a: float = 1e-8, # stability factor
name: str = None
)
Although, having an object, source and detector to simulate is in principle
enough to perform an FDTD simulation, One also needs to define a grid boundary
to prevent the fields to be reflected. One of those boundaries that can be
added to the grid is a Perfectly Matched Layer: or PML
. These
are basically absorbing boundaries.
# x boundaries
grid[0:10, :, :] = fdtd.PML(name="pml_xlow")
grid[-10:, :, :] = fdtd.PML(name="pml_xhigh")
# y boundaries
grid[:, 0:10, :] = fdtd.PML(name="pml_ylow")
grid[:, -10:, :] = fdtd.PML(name="pml_yhigh")
Grid Summary
A simple summary of the grid can be shown by printing out the grid:
print(grid)
Grid(shape=(161,97,1), grid_spacing=1.55e-07, courant_number=0.70)
sources:
LineSource(period=14, amplitude=1.0, phase_shift=0.0, name='source')
@ x=[48, ... , 51], y=[76, ... , 83], z=[0, ... , 0]
detectors:
LineDetector(name='detector')
@ x=[77, ... , 77], y=[0, ... , 96], z=[0, ... , 0]
boundaries:
PML(name='pml_xlow')
@ x=0:10, y=:, z=:
PML(name='pml_xhigh')
@ x=-10:, y=:, z=:
PML(name='pml_ylow')
@ x=:, y=0:10, z=:
PML(name='pml_yhigh')
@ x=:, y=-10:, z=:
objects:
Object(name='object')
@ x=11:32, y=30:84, z=0:1
Object(name=None)
@ x=84:116, y=32:52, z=0:1
Running a simulation
Running a simulation is as simple as using the grid.run
method.
grid.run(
total_time: Number,
progress_bar: bool = True
)
Just like for the lengths in the grid, the total_time
of the simulation
can be specified as an integer (number of time_steps
) or as a float (in
seconds).
grid.run(total_time=100)
Grid visualization
Let’s visualize the grid. This can be done with the grid.visualize
method:
# signature
grid.visualize(
grid,
x=None,
y=None,
z=None,
cmap="Blues",
pbcolor="C3",
pmlcolor=(0, 0, 0, 0.1),
objcolor=(1, 0, 0, 0.1),
srccolor="C0",
detcolor="C2",
show=True,
)
This method will by default visualize all objects in the grid, as well as the
field intensity at the current time_step
at a certain x
, y
OR z
-plane. By
setting show=False
, one can disable the immediate visualization of the
matplotlib image.
grid.visualize(z=0)
