FDTDImpulse - discrepancy between gprMax v.3.1.5 and v.3.1.7
Julina Hilgenberg
Hello everyone,
I’ve been working through the example from the YouTube tutorial “Day 3 - Impulse response & introduction to SFCW GPR simulation” by Dr. Antonis Giannopoulos.
The FDTDImpulse code works perfectly when I use gprMax v.3.1.5.
However, when I run the same code using the output files generated by gprMax v.3.1.7, the gaussiandotnorm trace and the impulse trace, which was convolved with a gaussiandotnorm, no longer match as well. The difference between the two traces is significantly larger than in the tutorial example.
The Hertzian dipole is used as the source in my simulations (for both the impulse and the gaussiandotnorm).
I noticed a difference in how the Hertzian dipole timesteps are handled between the two versions:
In v.3.1.5, the E-field generated by the dipole used half timesteps.
In v.3.1.7, E-field generated by the dipole now uses whole timesteps.
Could this change in timestep be the reason for the discrepancy in the FDTDImpulse output?
How could I adapt the FDTDImpulse example so that the gaussiandotnorm trace simulated with gprMax v3.1.7 better matches the impulse trace that has been convolved with gaussiandotnorm?
Any suggestions would be really appreciated!
Thank you.
Best regards,
Julina
