Vein modelling weirdness
JamesLally
Posts: 29
in Leapfrog
I have raised the following issue with Seequent Support and received a 'fix' of sorts but no explanation as to the cause and apparently no other users have reported anything similar. So I'd like to know if anyone else has come across this bug in vein modelling.
LF version 2021.1.3: I had a set of vein intercepts defined by a selection column in a merged interval table. I created a new geological model with Base Lithology set as the selection column. When I created a new vein with pinch-outs the resulting modelled shape passed through drill holes that didn't contain the corresponding vein code. Even if I coded the remainder (non-vein) parts of all drill holes with a dummy no vein code the model still pushed through drill holes it shouldn't have, although pinch-outs did occur in some holes.
The attached picture shows a blue modelled vein that should only be passing through drill holes containing the blue code and the circled area shows where the vein is modelled through holes where it doesn't occur. Note that in part of the model the pinch-out works as it should. I ran the same model with everything that wasn't the blue code assigned the same dummy 'no vein' code with the same result. This is one of 8 vein codes in my data and all did the same thing.
Now for the really weird part: when I re-saved the project and cut out some un-necessary parts to send to Support to have a look at, the resulting refreshed models seemed to fix themselves and behave properly. However, when I deleted and re-generated the models the same problem appeared again. The original model is still not working as it should, regardless of how many times I refresh or re-generate the models.
The support fix involved manually inserting polylines to control the hangingwall/footwall trends and force a pinch-out, but to me this is somewhat unsatisfactory and doesn't really explain what might be happening.
If anyone else has had this happen, please let me know along with any solution you found.
cheers
James Lally
LF version 2021.1.3: I had a set of vein intercepts defined by a selection column in a merged interval table. I created a new geological model with Base Lithology set as the selection column. When I created a new vein with pinch-outs the resulting modelled shape passed through drill holes that didn't contain the corresponding vein code. Even if I coded the remainder (non-vein) parts of all drill holes with a dummy no vein code the model still pushed through drill holes it shouldn't have, although pinch-outs did occur in some holes.
The attached picture shows a blue modelled vein that should only be passing through drill holes containing the blue code and the circled area shows where the vein is modelled through holes where it doesn't occur. Note that in part of the model the pinch-out works as it should. I ran the same model with everything that wasn't the blue code assigned the same dummy 'no vein' code with the same result. This is one of 8 vein codes in my data and all did the same thing.
Now for the really weird part: when I re-saved the project and cut out some un-necessary parts to send to Support to have a look at, the resulting refreshed models seemed to fix themselves and behave properly. However, when I deleted and re-generated the models the same problem appeared again. The original model is still not working as it should, regardless of how many times I refresh or re-generate the models.
The support fix involved manually inserting polylines to control the hangingwall/footwall trends and force a pinch-out, but to me this is somewhat unsatisfactory and doesn't really explain what might be happening.
If anyone else has had this happen, please let me know along with any solution you found.
cheers
James Lally
Tagged:
1
Comments
-
Hello James,I have also had the same problem with pinch-out vein modeling, however, I have found a way to solve the problem without adding polylines.It appears that vein modeling has this problem when the program needs more segments to interpolate at the edges of the vein. I show you an example in the attached figure.In this case I had the same problem and the solution was to add some extra segments at the edges of the grain to guide the interpolation. Another solution that has worked for me is to edit the HW or FW of the outer segments of the vein to control their shape so that this problem does not happen.Cheers0
-
Thanks for the response Kenneth. Can you clarify what you did in the example? it looks like you added an intercept into one of the drillholes in the bottom figure that's not in the top. Did you raise a support ticket for this by the way? From my interaction with Support it seemed liked no-one else has reported anything similar, which made me start questioning my sanity! There is presumably some particular confluence of factors causing the error, which is why it's not more widely reported. Otherwise it's just me.0
-
Hi @JamesLally and @KennethMontero,
Thanks for your comments.
James - from reading the details of your support ticket I understand you were shown a work-around.
We are well aware of this issue - it has been reported by a number of users. You are probably aware that we have been working on the vein tool for the last couple of releases.
The issue is encountered when you have a vein that contains 'natural' pinchouts - i.e. areas that are naturally pinched out because the footwall and hanging wall cross over. Where this occurs, drillholes outside of the 'naturally pinched-out vein' are excluded from subsequent pinchout generation.
This is the case with the RW_Deeps_S vein in the project you provided. In the image below, I show the 'naturally pinched out vein' and the pinch-out points generated when the 'pinch-out' option is ticked on. Note that these all fall inside the 'naturally pinched out vein'. The other holes are effectively ignored.
In the 2021.2.0 release (upcoming), we are providing an option to pinch veins out on un-logged traces (currently a hole must have a logged interval in it in order for a pinch-out to be generated). That change will not address the issue you have noted here though.
A second change, expected to be available in 2022.1 (~May), will force pinch-outs on all traces. This will directly address your issue. The image below shows an example of the vein generated when this option is selected, along with the pinch-out points generated. You may still want to constrain the vein with a boundary polyline, or force a pinchout with edits, in the areas not informed by drilling.
In the meantime, you'll need to use the edit tools to force the pinchout where you want it.
0 -
Thanks Mike.
I did get the workaround but found I couldn't actually replicate the same results on my own and went back to manually digitising the boundary I needed. When you say 'naturally pinched out' do you mean HW/FW crossovers that occur when vein modelling without 'pinchouts' turned on? If this is the case I am still confused because in my original model (and indeed a new model I have just created on the same drill database and project) if I don't select pinchouts the vein extends to the edges of the model through all drillholes.
I am also unsure about what you mean by 'forcing' a pinchout. I didn't think you could manually add new pinch-out points, only edit existing ones. Or are you referring to editing the HW and FW?0 -
Hi James,
"naturally pinched out" - yes exactly - where FW and HW cross over. This does occur in the vein model you provided when Pinchout is ticked off. But beware - you need make sure 'minimum thickness' is set to 0 (it was set at 1 in those veins). Minimum thickness becomes inactive when Pinchout is activated - but becomes active again if pinchout is deactivated. That will make the vein extend across the model.
Forcing a pinchout. You can't add to the set of pinch-out points created on drillholes. But you can add a psuedo pinchout with footwall and hanging wall polyline/point edtis. SImply add points to FW and HW that force the surfaces to cross.
Happy to have a chat and screen share if you would like.
0 -
Thanks for clarifying Mike. No need for further input, I will await next year's updates.0