Transfer fuction (S21) of whole GPR module (Tx,Rx antennas and the buried targets)

[Pamela Manase]

Dear gprMax community

I was kindly asking how to have access to the input source waveform in gprMax i.e in V/m directly from the Tx antenna.

I need to calculate the Transfer function (S21) of my whole model of GPR (Tx and Rx  antennas and the buried targets) 

==> To have an equivalent of Channel frequency response/ Impulse response of my whole GPR module from gprMax simulations

We have acess to  output from Rx ofcourse i.e in V/m

But I need the input information directly from the transmitter,  but I dont know how to acess it. I use herzian dipole + ricker waveform. 

Thank you so much in advance !

Best Regards

Pamela

[Antonis Giannopoulos]

Hi Pamela,

If you are modelling the antenna structure you should not really use the Hertzian Dipole source which adds a current density term in that point.  The transmission line model and voltage source is a better approach. The waveform that will be applied is given by the description (e.g. gaussian pulse) If you modelling the antenna structure a monopulse will work as the antennas kills the DC and in essence radiates a derivative. 

Another option is to see how the provided model of the 1.5GHz antenna is done especially in terms of the excitation.

Hope this makes sense! 

Best

Antonis

[Pamela Manase]

Dear Antonis

Thank you so much for your reply. I understood , will take a closer look into a correct type of source to use for my model.

Regards

Pamela

[Pamela Manase]

Dear Antonis

Is there any means that I can access the Electric field emitted in V/m directly the source i.e from voltage/transmission line source + ricker waveform?

Just to be more clear:

Example if I am using voltage source+ricker ==> amplitude of source waveform will be in volts, then  I thought of  using this formula V=-Ed  to calculate the emitted field from the transmitter in V/m ?

Aim

I need to have the Channel Frequency Response H(f) OR Channel Impulse Response  h(\tau)  of the whole GPR system model (please see the attachment) I have design in gprMax (voltage source + buried target + Rx )

        H(f) = E_out(f) / E_in(f)

E_out (f) ==> I can have it from Rx ==>as a matrix ==>  field(Ntime x Nposition)

E_in(f) => This is where I am stuck on how to access this data from gprMax ( I do not prefer to use direct wave since this value is recorded from Rx and involve coupling with the ground and it is accesed from Rx not Tx)

Thank you so much in advance, I am sory for the long explanation !!

Regards

Pamela

[Antonis Giannopoulos]

Hi Pamela,

In FDTD Voltage across a Yee cell edge will be simply related to Electric field as V=-E*DL In essence this is just a scaling factor and not as important as far as you are consistent on what you are doing. In exciting a model there are a number of options. 1. to add a current density term at a point and that creates a "transparent no-reflecting source" (i.e. incoming waves do not get reflected by its presence) and this is how you get a Hertzian Dipole or 2. to add a Voltage Source with or with out resistance using a similar analogy by creating a small circuit relation at this cell edge. We do this for the Voltage source formulation with resistance. If you use an old style hard source you then prescribe the electric field in this location and it is called a hard source because when it goes to zero it is equivalent to placing a small PEC edge there (E =0) and can reflect incoming waves. 3. The transmission line source tries to do something similar to the Voltage source by connecting a 1D FDTD to the 3D one at the excitation point offering both reflected and incident pulses. It is an approximation that works better for monopoles (coaxial feeds but we do not implement this yet) but it is OK for the dipole antennas we often use but not perfect (this is another story ...)

What it seems you want to do is to deconvolve from your response the "forcing function" (i.e. the source) to obtain the impulse response of the system. As far as the amplitude scaling of both receiver and transmitter are the same it will work. The other way is to excite the model with an impulse (1,0 ...) and this will give you the "digital FDTD Green's function" of the model. Now this response is not valid as it is as the FDTD solution is not valid for all wavelengths. However, if you convolve this response with any properly weighed in terms of its spectrum excitation function the results will be identical with the ones that you will get if you had excited the model with this function in the first place. Alternatively you can FFT the "erroneous" impulse response which will give you in the frequency domain the H(f) but this will only be valid for the frequencies that the simulation is accurate (i.e. the ones that the Lambda/10 criteria is satisfied) these frequencies are a lot less than the ones contained by the transformation that goes up to the Nyquist one. 

Obviously, if this all happens in 2D the discussion is practically academic as it will not really relate to a proper system function for obvious reasons ...

Hope this helps (or I might not have understood your questions ...)

Best

Antonis

[C Master]

Hi  Pamela Manase ，

I find that my problem is similar to yours, and I hope your problem has been solved.

 I'm going to use the Hertzian dipole as the source, but I don't know how to set its current value. I have a fixed power for this source, and how do I use that power to calculate the current from this source, or the strength of the electric field that the source is emitting.

best，

C Master

[Antonis Giannopoulos]

It is easy to see in the documentation how the Hertzian Dipole is implemented

http://docs.gprmax.com/en/latest/input.html#hertzian-dipole

When you use the Hertzian Dipole source your waveform specifies the time variation of the current pulse on it I(t) in Amperes. 

Best

Antonis