Create a 3D layer model about gimli HOT 9 OPEN

Dear colleague,
when raising an issue, there is a template that should be filled describing some background, what you did, what you expected and what actually happened. You are copying your code (not even into the part where it should go) along with chinese symbols, no images, but an error message that is clear: You do not have a region with marker 1.

from gimli.

You should use only one call of mt.createWorld(). There you can specify layers by the layers argument, see
https://www.pygimli.org/pygimliapi/_generated/pygimli.meshtools.html#pygimli.meshtools.createWorld
and the below linked examples (mostly 2D) using this function.

from gimli.

"Sorry, my explanation was not very clear."

"I want to generate a 3D layered model and set an anomaly in it to depict the vertical membrane structures in a landfill. During inversion, I want to keep the resistivity of the background model constant and only invert the central vertical membrane area. However, I'm not sure how to set up the 3D layered model, as it seems the 'layer' parameter can't be used in three dimensions. I'm encountering the following error messages in my code."

from gimli.

Have you tried mt.createWorld(start=[-10, -10, -40], end=[140, 10, 0], layers=[-10, -30])?

Note that you will need a much larger mesh for accurate simulations.

from gimli.

"It displays an error like this,"

from gimli.

Yes, I am sorry, obviously this feature does not exist yet. I would suggest the following:

• creating a world as written but without layers (marker 1)
• adding a cube as a middle layer (marker 2)
• setting a region marker (marker 3) into the first layer

from gimli.

from pygimli.physics import ert
import pygimli as pg
import numpy as np
import matplotlib.pyplot as plt
import pygimli.meshtools as mt
world=mt.createWorld(start=[-10, -10,-40], end=[140, 10,0],worldMarker=True,marker=1)
cube3=mt.createCube(size=[128, 0.4, 40], pos=[128/2, 0, -40/2],marker=4)
start_model = world +cube3
pg.show(start_model,alpha=0.3,markers=True)

filename = "cc.shm"
shm = pg.DataContainerERT(filename)
data=ert.load(filename)
data['k']=ert.geometricFactors(data)
data['rhoa']=data['k']*data['u']/data['i']*1000
data.remove(data['rhoa']<0)
data['err'] = ert.estimateError(data,relativeError=0.02)
for s in shm.sensors():
start_model.createNode(s)
for s in shm.sensorPositions():
start_model.createNode(s - [0,0,1e-2/2])
inversion_mesh= mt.createMesh(start_model,quality=1.4)
pg.show(inversion_mesh, markers=True, showMesh=True)
mgr=ert.ERTManager()
mgr.setData(data)
mgr.setMesh(inversion_mesh)
mgr.inv.setRegularization(1, background=True)
mgr.inv.setRegularization(1,fix=200)
mgr.inv.setRegularization(2,startModel=1e4,limits=[100,1e4+20])
mgr.invert(lam=10,zWeight=0.1,verbose=True)
mgr.saveResult()
mgr.showMisfit()

I have created a model that includes an anomaly. In the inversion process, I want to constrain the resistivity of the background to remain constant, optimizing only the resistivity of the anomalous cube. Now, I want to change the uniform background model to a layered model, assigning different resistivities to each layer, and similarly constrain the resistivities of the layered background to remain unchanged during inversion

When I haven't established a layered structure, my forward modeling mesh is correct, as shown below

However, when I try to use a layered model, there is a problem with establishing the mesh; the mesh is not divided correctly

from gimli.

cc.zip

from gimli.

Please paste code and messages as text instead of screenshots.

In your case the layers (cubes) intersect with the vertical sheet. Whereas such things are automatically resolved in 2D, in 3D this needs to be solved by splitting every cube into two, one on either side of the sheet.

from gimli.