Hi, I'm a italian geologist and I’m starting to use Dinver’s HVSR data inversion for professional reasons. I’ve installed Geopsy 3.3.6 version and, once I loaded the* .hv file and setted the layers parameters, when I start the inversion process, the Itmax string and seed string doesn’t appear in the runs window and therefore the inversion process doesn’t start. I tried to reinstall the program but the result doesn't change. Someone can help me!
Luca Laveni
Dinver inversion problem
-
- Posts: 2
- Joined: Wed Sep 29, 2021 7:58 am
- Location: Desio (MB) Italy
- Contact:
Re: Dinver inversion problem
Hi Luca,
Itmax and seed were present in old releases. A user-specified seed is meaningless for a parallelized job. It is used only for debug when starting dinver option "-j 1" (only one thread). In such case a seed can be specified from the command line, but it is removed from the graphical interface. Itmax is now integrated in Ns. Ns is now the total number of models to generate. In the past, it was the number of models generated at each iteration so that the total number of models was itmax*Ns.
Hence this is fine if you do not have these fields. There is no need to re-install.
Be careful that dinver tries to invert the H/V curve assuming that it is a pure Rayleigh wave which is almost never the case for experimental data processed with the classical H/V processing. The Love wave and the body wave contributions can significantly increase the H/V amplitude. More advanced techniques such as Raydec (Hobiger et al. 2009), HVTFA (more info in Neries JRA4) for single station processing can filter out these waves. From array processing, the Rayleigh ellipticity can be obtained by Poggi et al 2010, WaveDec (Marano et al 2012) or RTBF (Wathelet et al.2018).
Best regards,
Marc
Itmax and seed were present in old releases. A user-specified seed is meaningless for a parallelized job. It is used only for debug when starting dinver option "-j 1" (only one thread). In such case a seed can be specified from the command line, but it is removed from the graphical interface. Itmax is now integrated in Ns. Ns is now the total number of models to generate. In the past, it was the number of models generated at each iteration so that the total number of models was itmax*Ns.
Hence this is fine if you do not have these fields. There is no need to re-install.
Be careful that dinver tries to invert the H/V curve assuming that it is a pure Rayleigh wave which is almost never the case for experimental data processed with the classical H/V processing. The Love wave and the body wave contributions can significantly increase the H/V amplitude. More advanced techniques such as Raydec (Hobiger et al. 2009), HVTFA (more info in Neries JRA4) for single station processing can filter out these waves. From array processing, the Rayleigh ellipticity can be obtained by Poggi et al 2010, WaveDec (Marano et al 2012) or RTBF (Wathelet et al.2018).
Best regards,
Marc
-
- Posts: 2
- Joined: Wed Sep 29, 2021 7:58 am
- Location: Desio (MB) Italy
- Contact:
Re: Dinver inversion problem
Thanks a lot Marc!
Luca Laveni
Luca Laveni