Extended Model for Resistivity in Pygimli

[tonapawilliam]

Excuse Me Sir, Is there a function in Pygimli ERT to output the model inversion results in the form of an extended model like in res2dinv, so that the output data can be shaped like an inverted trapezoid or triangle? If so, can the data be exported in accordance with the shape of the extended inversion model only? Thank you

[halbmy]

In res2dinv, the model is trapezoidal-shaped which can be unfavorable for certain conductivity distributions and with topography in general. Of course you can easily create such a trapezoidal model shape in pyGIMLi as well, maybe it could be beneficial to add such an option to meshtools.createParaMeshPLC.

However, you can blend out poorly covered cells by using

mgr.showResult(coverage=mgr.standardizedCoverage(threshold))

e.g. for the gallery (2d flat topo example) case, mgr.showResult(coverage=mgr.standardizedCoverage(-1.3)) leads to

[halbmy]

Just for fun, I just added (8050bef3) a keyword argument trapRatio to createParaMeshPLC that might help to reduce the number unknowns and also the ill-posedness of the inverse problem. The example

mgr = ert.Manager(data)
mgr.invert(trapRatio=0.3, paraDepth=12)
mgr.showResult(coverage=1)

yields

[tonapawilliam]

Thank You Sir. When I export this data into x y res, can this trapezoidal model also come out so that when plotted based on x y, it will also form a trapezoid? I find it quite problematic when I take out all the data, including the extrapolated ones that cause my inversion data to reach more than 7000.

Edit: I have tried it, but it seems like the data remains the same amount and when x and y are plotted, it remains square. Is there a function for that?

I also tried to make a trapezodial, but somehow I couldn't do it, and instead the result is as follows.

Can you help me edit my code for this? Or tell me where my code is missing AreaNew.txt

[halbmy]

For using the trapRatio argument, you will have to checkout pyGIMLi by git as described by https://www.pygimli.org/installation.html#staying-up-to-date

[tonapawilliam]

OK, Thank You Very Much Sir