How to solve the polarization problem

[longxiang yao]

Dear GPRMAX team,

The figure below presents the four set of data I acquired from the same 3D GPRMAX model. The sole distinction lies in the polarization mode. It is distinctly observable that both the xz and zx polarization exhibit stronger direct waves than the co-polarization (xx and zz), see figure 2.

Q①: Is this due to the antenna distance issue in the GPRMAX command settings, or is it because of the different electric field intensities at the antenna endpoints and sides?

And after background removal, the results are shown in figure 3.

Q②: In our initial hypothesis, we anticipated that for polarization, the effect of co-polarization would be more prominent than that of cross-polarization. However, my results indicate the contrary. I'm curious to know what might be the underlying cause.

I'm very much looking forward to your reply.

[liu ft]

Hello, 

    I am also experiencing the same issue. Have you resolved it? 

I'm very much looking forward to your reply.

[Antonis Giannopoulos]

Your source is on the dielectric interface and  it will be better if it is a few cells further up. In reality your actual antenna structure is not on the interface. So, your y coordinate for the source and the receiver can be 0.758 this puts your source and receiver in air. It is still close to the interface but it is more realistic. The other issue you have is that in FDTD field components are not at the same location and as a result your cross polarisation component is not exactly at the same x position as the source. To get a better simulation of this you can take the average of two z components by getting the output also at z at 0.048 and then average the z components. 

So, your source and receiver should be

#waveform: ricker 1 12e8 my_ricker 

#hertzian_dipole: x 0.3 0.758 0.05 my_ricker
#rx: 0.4 0.758 0.05

#rx: 0.4 0.758 0.048

Try this and let me know how it worked. 

Also, in gprMAx models instead of defining the material "air" you can use "free_space" without definition as well as "pec" as they are already known in the code.

Antonis