This week in SimPEG…
Thibaut Astic has been working on petrophysically guided inversions. Our petrophysically guided scheme assumes a Gaussian distribution of the physical properties, in some space, for each distinguishable contrast/unit. To tame this assumption, we push the concept of
mapping so that each physical property and each rock unit has its own
mapping. This allows us to enforce in the inversion nonlinear petrophysical relationships, different from one rock unit to the other. This example is simply a proof of concept, using a joint
inversion of 1D linear
problems with 3 distinct units, each with a different
mapping (identity, quadratic and cubic). Notice how the petrophysical relationships can be well recovered in this simple example (top row).
This week, Doug Oldenburg and I (Lindsey Heagy) submitted a paper on simulating DC, FDEM and TDEM on 2D and 3D cylindrical meshes to Computers and Geosciences. The examples revisit some of the formative papers on steel cased wells and electromagnetics. All of the notebooks are available at: https://github.com/simpeg-research/heagy_2018_emcyl, and the preprint is available on the ArXiv (https://arxiv.org/abs/1804.07991). Thanks especially to Thibaut Astic and Dikun Yang who read the whole thing (it is long!) and gave valuable feedback.
Rowan Cockett shared some good news… the paper on Richards equation was accepted: https://www.sciencedirect.com/science/article/pii/S0098300417311494