SimPEG 0.15.0 Release Notes#

May 19th, 2021

This minor release encompasses many different additions and improvements that have been made to SimPEG since the previous release!



We have merged @thast’s petrophysically guided inversion (PGI) methodology into SimPEG! There any many new directives and objective functions related to this methodology of joint inversion. There are also now new tutorials showing how you can implement this in both joint inversions and single-domain inversions to incorporate petrophysical information into the inversion! Check out the multiple PR’s related to this branch addition for more details.

  • #937: Improving BetaEstimate_byEig (and adding other directives for the other terms of the objective function)

  • #938: add the Wolfe Curvature condition to the optimization

  • #939: PGI

  • #959: Tutorial: Joint PGI Grav + Mag + Octree

This is a great first addition for SimPEG’s joint inversion capability and we will be adding additional forms soon!

Line Current Source#

After a long wait, and multiple testing issues, we were finally able to merge in the frequency domain line current source branch. This branch also enables the line current source to be used on TreeMeshs.

Boundary Conditions#

With the most recent 0.7.0 version of discretize, we have updated the DC boundary conditions to flexibly support more mesh types. As part of this, we have also changed the default boundary condition for the simulations in SimPEG.electromagnetics.static.resistivity to be the Robin type condition (equivalent to the Mixed option previously). This enables both Pole-Pole solutions for the nodal formulation (something that was not possible previously). We also caught a error in the 2D DC resistivity simulations’ previous Mixed boundary condition option that caused incorrect results.

Static Utils#

The static_utils module within SimPEG.electromagnetics.static.resistivity has been given a pass through with many internal changes to make it more flexible in its handling of dc resitivity surveys. Wenner-type arrays should be better supported by the psuedosection plotting utility. It now also includes a 3D psuedosection plotting command using plotly (which should display as intractable within web pages).

Future Warnings#

We have changed all of the DeprecationWarning from 0.14.0 to FutureWarning. We have done a pass to double check that all internal calls should not issue these warnings, but let this serve as a notice that the items that issue a FutureWarning will be removed in the next minor release of SimPEG (0.16.0)

Bug Fixes#

There are also many individual bug fixes that went into this related to many different items. For the complete updates, check out the list of PR’s that have been merged into SimPEG.


This is a combination of contributors and reviewers who’ve made contributions towards this release (in no particular order).

Pull requests#

  • #791: FDEM line source

  • #932: Load ubc dcip datatypes and plot 3d pseudosections

  • #937: Improving BetaEstimate_byEig (and adding other directives for the other terms of the objective function)

  • #938: add the Wolfe Curvature condition to the optimization

  • #939: PGI

  • #958: Mag. Simulation: fix if statement for self.chi

  • #959: Tutorial: Joint PGI Grav + Mag + Octree

  • #960: Fix IP simulation when model or field is None

  • #964: Replace Travis Icon by Azure in Readme

  • #965: Add conda-forge badge

  • #967: Bug fix for gravity simulation

  • #969: Fix dc2d: Pseudosection and add warning in examples for the observation file format

  • #974: Clean up and refactor

  • #977: Fixing 2D IP bug

  • #982: Update

  • #983: Update README.rst

  • #984: typo in the init

  • #986: Update README.rst

  • #987: Ensure the 2D IP simulation stores it’s scale array as a float

  • #993: Dc boundary conditions updates

  • #994: Update

  • #997: Update api_mag.rst

  • #999: Update

  • #1002: typo

  • #1003: Remove the duplicate code

  • #1005: 0.15.0 release

Closed issues#

  • #766

  • #919

  • #935

  • #955

  • #976

  • #980

  • #992

  • #998

  • And many others without issues.