Hi everybody,
For the DC resistivity example, can I change configurations including the position of the sources and receivers to simulate for example a borehole-to-borehole case and the mesh with a Tensor mesh instead of an OcTree mesh? Thanks.
Hi Astor,
Yes, you can place the electrodes as you wish. You need to create a survey object with the electrodes layout as you wish.
See for example this tutorial that show the creation of a survey object from ABMN locations: http://docs.simpeg.xyz/content/tutorials/05-dcr/plot_inv_2_dcr2d.html#sphx-glr-content-tutorials-05-dcr-plot-inv-2-dcr2d-py
Hi Thibaut,
Thank you for your reply and tutorial. I have figured out the configuration and now I am trying to use SimPEG to do a joint inversion. Could you please tell me whether there are tutorials of “directives”? I didn’t find one on the Website. Thanks!
Indeed I do not think we have tutorials for directives. It’s a bit harder as their goals are often to serve problem-specifying purposes. So we document them more through examples. The gallery is the place to look for that. What do you want to do with directives?
I’d like to achieve the parameter adjustment for the joint inversion in a similar way you achieved on SimPEG with directives. For joint inversion, there are more parameters than single physics inversion, such as beta1, beta2 for different physical models. I am not quite clear on how to implement this with directives. You mentioned that there are examples. Is there anyone that is related to the joint inversion? I saw the examples in the DC and travel time inversion. They are quite straightforward. But for the joint inversion, I think there may be some differences.
Disclaimer: Joint Inversion is something that people are doing with SimPEG at the moment but at a research, academic levels, and as such these implementations are not part of the main SimPEG distribution yet. Several initiatives are currently ongoing for bringing coupling terms such as petrophysical or structural coupling within the main distribution but it will be some times before there are in.
As such, there is no directives specifically related to joint inversion yet. We have examples on how the already available directives can be used. For your purpose, you will likely have to write your own directives. A good start might be to look at the code of the BetaSchedule directives, which controls the beta-cooling, and adapt it into a directive that fulfill your purpose.
You will likely want to work from your own clone of the SimPEG Github repository. If this is something you are not familiar with, you can check this introduction: http://docs.simpeg.xyz/content/basic/getting_started_developers.html
Directives are stored under <SimPEG repository>/SimPEG/directives.py
If you ran into specific issues or have questions while implementing, we will be happy to help you with those as well.
Thank you very much for your reply and explanation. I will start with the BetaSchedule directives.
Hi Thibaut,
I encountered a problem about the spatial differential operator when trying to achieve the joint inversion. I posted a question in the “discretization” module but there is no response. Could you please take a look if you have time? I really appreciate your time and help. Thank you very much.