Using the NF2FF with a mirror plane. Correct usage help needed.

[Aspire-Design-Ltd]

Hi. I am trying to simulate a quarterwave monopole over a large ground plane. The classic radiation pattern is shown on wikipedia as:

Using OpenEMS to simulate a 1/4wave monopole on a 10x10 wavelength ground plane with a mirror plane the farfield calcs do not come anywhere near to the classic radiation pattern above.

The NF2FF code allows a mirror plane to be positioned on any axis and at any mesh line. The octave code used is:
nf2ff = CalcNF2FF(nf2ff, Sim_Path, f0, thetaRange*pi/180, phiRange*pi/180, 'Center', [0,0,port_length],'Mirror',{2, 'PEC', "0.0"});
This puts a PEC mirror plane on the 'z' axis at position 0. The notation to define the mirror plane is:
{ direction, type, position }

• direction -> axis 0 -> x, 1 -> y, 2 -> z
• type -> "PEC", "PMC"
• position -> mesh line position for the mirror plane. Must be a float.

The function CalcNF2FF adds the following to the 'nf2ff.xml' file for the NF2FF processing code.
<Mirror Dir="2" Type="PEC" Pos="0.0"> </Mirror>
Using paraview the results for the farfield do not look correct. Below shows the farfield for the z-axis mirror plane at 0.0 [LEFT image] which is where the PEC boundary is and also shown is the mirror plane moved to position 2.0 [RIGHT image] which is the antenna bottom. The wire frame shows the orientation of the FDTD volume.

Anybody got any thoughts??? thanks.

FURTHER WORK:
I tried extending the Z-Axis out 5 wavelengths in-case the far field calcs need a more symmetrical FDTD volume. Also put the centre of the NF2FF calculation to the middle of the quarterwave monopole but still no real progress.
Below is the code and radiation pattern so far.
Monopole4.zip

[VolkerMuehlhaus]

I think you are taking too many steps at once, and suggest that you start from a known-good antenna example.

The concern with pattern calculation for PEC backside was raised by me. It's based on one model for an RFIC 240 GHz patch antenna with PEC backside ground. I didn't spend the time to troubleshoot that, and changed my model to have PML all around.

For your case, you can also try to explictely model the ground plane, with PML8 all around, and see if that looks reasonable. This is just to ensure that there are no other issues in your model code or evaluation method.

I would also suggest to use the "standard" code from other openEMS antenna models to get the 2D cut planes, to avoid any issues with your pattern evaluation in 3rd party tools. My advice is to go step by step when troubleshooting this, you had already seen with the AddWire() vs. AddCylinder() that it's easy to misunderstand things.

I can't help with more specific code because I dropped the Matlab/Octave models in favour of Python openEMS models.

[VolkerMuehlhaus]

I don't know if this helps, but here is a Python (not Octave!) monopole example that I created for you from the patch antenna example. This is with explicit (drawn) ground plane, the next step would be to replace the ground plane with PEC boundary condition.

# monopole, derived from patch antenna example

### Import Libraries
import os, tempfile
from pylab import *

from CSXCAD  import ContinuousStructure
from openEMS import openEMS
from openEMS.physical_constants import *


# change current path to model script path
os.chdir(os.path.dirname(os.path.abspath(__file__)))
Sim_Path = os.path.join(os.getcwd(), 'monopole_data')
print("Sim Path: ", Sim_Path)


### General parameter setup
post_proc_only = False
unit = 1e-3  # define drawing unit mm
f_target = 2.4e9  # target frequency

# monopole dimensions
wavelength_air = 3e8/f_target/unit  # in drawing units
antenna_length = wavelength_air/4  
antenna_diameter = 1 

groundplane_size = 2*wavelength_air
margins = wavelength_air/2  # simulation margin around ground plane

#setup port parameters
feed_length = 1 # port size in drawing units
feed_R = 50     # feed resistance

# size of the simulation box
SimBox = np.array([groundplane_size+2*margins, groundplane_size+2*margins, 2*antenna_length+2*margins])

# setup FDTD parameter & excitation function
f0 = f_target # center frequency
fc = 1e9 # 20 dB corner frequency

### FDTD setup
## * Limit the simulation to 100k timesteps
## * Define a reduced end criteria of -40dB
FDTD = openEMS(NrTS=100000, EndCriteria=1e-4)
FDTD.SetGaussExcite( f0, fc )
FDTD.SetBoundaryCond( ['MUR', 'MUR', 'MUR', 'MUR', 'MUR', 'MUR'] )

CSX = ContinuousStructure()
FDTD.SetCSX(CSX)
mesh = CSX.GetGrid()
mesh.SetDeltaUnit(unit)
mesh_res = wavelength_air/20

### Generate properties, primitives and mesh-grid
#initialize the mesh with the "air-box" dimensions
mesh.AddLine('x', [-SimBox[0]/2, SimBox[0]/2])
mesh.AddLine('y', [-SimBox[1]/2, SimBox[1]/2])
mesh.AddLine('z', [-SimBox[2]/2, SimBox[2]/2])

# create port
start = [-antenna_diameter/2, -antenna_diameter/2, 0]
stop  = [ antenna_diameter/2,  antenna_diameter/2, feed_length]
port = FDTD.AddLumpedPort(1, feed_R, start, stop, 'z', 1.0, priority=5, edges2grid='xy')

# create monopole
antenna = CSX.AddMetal( 'antenna' ) # create a perfect electric conductor (PEC)
start = [-antenna_diameter/2, -antenna_diameter/2, feed_length]
stop  = [ antenna_diameter/2,  antenna_diameter/2, antenna_length]
antenna.AddBox(priority=10, start=start, stop=stop) # add a box-primitive to the metal property 'patch'
FDTD.AddEdges2Grid(dirs='xy', properties=antenna, metal_edge_res=mesh_res/2)

# create ground plane
gnd = CSX.AddMetal( 'gnd' ) # create a perfect electric conductor (PEC)
start = [-groundplane_size/2, -groundplane_size/2, 0]
stop  = [ groundplane_size/2,  groundplane_size/2, 0]
gnd.AddBox(start, stop, priority=10)

FDTD.AddEdges2Grid(dirs='xy', properties=gnd)

# add extra mesh cells around monopole
mesh.AddLine('x', [-antenna_diameter, antenna_diameter])
mesh.AddLine('y', [-antenna_diameter, antenna_diameter])

# add extra cells to discretize the monopole length and port size
mesh.AddLine('z', [0, feed_length/2, feed_length, antenna_length-antenna_diameter])
mesh.AddLine('z', linspace(feed_length, antenna_length, 10))

mesh.SmoothMeshLines('all', mesh_res, 1.4)

# Add the nf2ff recording box
nf2ff = FDTD.CreateNF2FFBox()

### Run the simulation
if 1:  # debugging only
    CSX_file = os.path.join(Sim_Path, 'monopole.xml')
    if not os.path.exists(Sim_Path):
        os.mkdir(Sim_Path)
    CSX.Write2XML(CSX_file)
    from CSXCAD import AppCSXCAD_BIN
    os.system(AppCSXCAD_BIN + ' "{}"'.format(CSX_file))

if not post_proc_only:
    FDTD.Run(Sim_Path, verbose=3, cleanup=True)


### Post-processing and plotting
f = np.linspace(max(1e9,f0-fc),f0+fc,401)
port.CalcPort(Sim_Path, f)
s11 = port.uf_ref/port.uf_inc
s11_dB = 20.0*np.log10(np.abs(s11))
figure()
plot(f/1e9, s11_dB, 'k-', linewidth=2, label='$S_{11}$')
grid()
legend()
ylabel('S-Parameter (dB)')
xlabel('Frequency (GHz)')

idx = np.where((s11_dB<-10) & (s11_dB==np.min(s11_dB)))[0]
if not len(idx)==1:
    print('No resonance frequency found for far-field calulation')
else:
    f_res = f[idx[0]]
    theta = np.arange(-90.0, 90.0, 2.0)
    phi   = [0., 90.]
    nf2ff_res = nf2ff.CalcNF2FF(Sim_Path, f_res, theta, phi, center=[0,0,1e-3])

    figure()
    E_norm = 20.0*np.log10(nf2ff_res.E_norm[0]/np.max(nf2ff_res.E_norm[0])) + nf2ff_res.Dmax[0]
    plot(theta, np.squeeze(E_norm[:,0]), 'k-', linewidth=2, label='xz-plane')
    plot(theta, np.squeeze(E_norm[:,1]), 'r--', linewidth=2, label='yz-plane')
    grid()
    ylabel('Directivity (dBi)')
    xlabel('Theta (deg)')
    title('Frequency: {} GHz'.format(f_res/1e9))
    legend()

Zin = port.uf_tot/port.if_tot
figure()
plot(f/1e9, np.real(Zin), 'k-', linewidth=2, label='$\Re\{Z_{in}\}$')
plot(f/1e9, np.imag(Zin), 'r--', linewidth=2, label='$\Im\{Z_{in}\}$')
grid()
legend()
ylabel('Zin (Ohm)')
xlabel('Frequency (GHz)')

show()

[Aspire-Design-Ltd]

Thankyou @VolkerMuehlhaus your thoughts which helped me solve the issue. Reading some of the research articles into FDTD using a mirror planes I realised that its wrong to blindly just put the NF2FF surface box close to the edges of the simulated FDTD volume. This is especially the case if you plan to use a mirror plane.
What I did was extend the FDTD volume to 3 wavelengths with the PEC ground plane but I set the NF2FF surface, where the readings for the EM fields are recorded, at around a quarter wavelength from the antenna. In this way the simulated FDTD volume is playing out the EM fields for a quarter wavelength monopole in a large volume but for NF2FF calcs the surface/volume much closer to the antenna.
New images are now looking more like the classic.

Octave file : Monopole5.zip