First of all the news: After 1 1/2 years we can celebrate the release of a new AthosGEO version - 2.3.0!
- Users of the free AthosGEO View version (current and new ones) can register on the website, download and install it.
- Users with a license for AthosGEO Blend will be able to do the same, downloading from the customer section.
Note that the new version will not replace an existing installation: If you want, you can keep it - "just in case" -, and only remove it once you want to definitely switch to the new version and want to release the disk space.
Also if you have an activated license (AthosGEO Blend), you do not have to re-activate after the installation: It will be automatically recognized, and both installed versions will be fully functional.
AthosGEO View
Users of the free AthosGEO View variant will mostly benefit from the updates in the underlying ParaView only: We migrated from version 5.10 to 5.13, and these are the release notes for these 3 steps ahead:
- ParaView 5.11.0 Release Notes
- ParaView 5.12 Brings Exciting New Features
- ParaView 5.13.0 Release Notes
Since ParaView is a general purpose 3D visualizer for large to huge data sets, some of the updates may be more and others less interesting for geo data visualization. However, in any case it is worth having a look into the above release notes!
AthosGEO Blend
Most of the programming for the new version was done for the cement specific AthosGEO Blend variant. Just a few highlights:
- A User's Manual - finally! So far users had to rely on their notes, short "help" for the different "filters", a number of short presentations about single subjects and own experience; only for ParaView and AthosGEO View manuals did exist.
- A filter that checks a block model for inconsistencies that would lead to malfunctioning of the "Scheduler" module, like "holes" in the model, overlaps between blocks, sub- or superblocking (which is not working with the Scheduler) etc.
- A new "waste type" - besides the already existing one. This is the most complex update, and it deserves a broader explanation.
Two Waste Types
Most of the users would probably love to have a one-button function that handles wasting fully automatically. Well - this is not what you are getting! Rather your own expertise and reasoning will be even more challenged because you will get many more options and variables that you can and have to adapt. The good thing is of course the fact that you gain many more ways to adapt wasting to what you really want to achieve.
Let's start with a short explanation about how the "Scheduler" module is working. First of all it's job is of course to produce cement raw mix, one or several products, according to your quality and other specifications. It will also respect certain mining angle constraints and directional preferences, but for now we will leave these things aside.
In order to produce, often wasting is required as well, so this becomes the second job of the "Scheduler". But while we have quality and other specificytions for the products - what would be the specifications for "waste"? Sure also here we apply some quality constraints - mostly because there are materials in the quarry that we certainly do not want to waste. But first of all, wasting should be guided by the need to open a way for further production.
Wasting thus basically has to "know" where this further production will eventually happen. Meaning: it needs to have some look-ahead facility - a bit like a chess or other gaming program. This comparison is not fully correct of course, but it nicely illustrates the fact that a "perfect" solution of the problem would require an incredible amount of computation power and memory because of the sheer number of possible ways ahead! Instead we are addressing the problem from two sides with not 100% "perfect", but still very much viable solutions:
1. Short Term Waste
Look ahead at the next production, but only one single step: If production is not any more possible at some point, do some wasting instead and see if production can continue after that. Not more! But turns out that already with this, quite a number of wasting situations could be resolved. This is the wasting strategy that was implemented in AthosGEO Blend already from the beginning and it makes no sense to drop it: We only rename it from just "Wasting" to "Short Term Wasting" because this is what it actually does. But this is also why it fails sometimes: It postpones wasting as much as possible. Every quarry manager understands what can be the consequence: At some point so much wasting is needed for further production that one single wasting step is simply not enough!
2. Strategic Waste
But this problem can be solved because in a situation like the above it is often enough pretty obvious from the geological situation already what the problem is: "We should have started with overburden stripping much earlier", "In order to come to this lower limestone layer, we should have started to remove some of the sandstone above", etc. etc.
The new "Strategic Waste" feature simply makes use of this knowledge. It is defined mostly like just another product - only that it will not be counted as such for the production tonnage. Instead, the user defines certain block constraints (like "must be overburden") and a waste ratio range for that specific "strategic waste product". The following example will clarify the two concepts a bit more.
Above (at the top) you see a very little synthetic example model, with some high grade limestone, a Mg limestone and a marl unit, plus others that are not suitable for cement production. Very obviously, overburden and part of the bedrock (shale) will have to be wasted - the latter in order to generate a non-vertical mining angle.
The first step will always be an overall analysis using the "Optimizer" module. See to the left a screenshot of the same synthetic model as above, now with the result of an "Optimizer" run: In red those blocks are marked that contribute to the theoretical maximum of raw mix production from this quarry (together with some iron corrective and sand): 3129 ktons (or 3.1 mio tons) - good enough for running for up to 10 years a very small cement plant that needs only about 300 ktons per year only.
Now it remains to be shown how many of these "up to 10 years" can be realized, and how. This is now the job of the "Scheduler". The following screenshots are showing a number of variants that were generated with it. But please keep in mind that these are just examples - with many more strategy variants still being possible!
This is a variant that does not use the new "strategic waste" feature yet, but only the existing "short term waste". As you can see, it nicely does some stripping and also removes a certain quantity of the marl of which we need only a part, uncovering the underlying limestone.
The sequence of screenshots is always one row for one period - typically one year -, and shows "taken" blocks in red (production of raw mix - left), "wasted" blocks (middle) and finally the quarry with it's geological units after removing production and waste for that specific period (right).
But wasting only happens "last moment", and after finishing about 1/4 of the 5th period, a situation has come that one single wasting step was not any more enabling more production, although there is obviously still a lot of material available. In the waste ratio chart (at the bottom) you can see how initially no wasting was done, because of no "emergency". And during the last period (that was only partially finished), the waste ratio was "desperately" raised, but this was not sufficient. The total production after all achieved periods is 1272 ktons.
The lower right screenshot illustrates very nicely the reason why the "Scheduler" was terminating. You have to know that for the example, a mining angle of 45° was chosen in all directions that must not be exceeded, and you need to realize that you will always need at least a little amount of the "Marl" unit (yellow), together with one of the two limestones (dark blue and pink), but
- you can neither do a "wasting" step except for "Overburden" which does not uncover these units in a way that they can be taken without exceeding the mining angle. "Mg Limestone" cannot be wasted because it is all covered with "Overburden" (it would require a second wasting step!), "Marl" cannot be wasted because the angle would become too steep, and "High Grade Limestone" must not be wasted anyway because it is excluded in the settings (it is too valuable for the production!).
- nor can you do a "production" step because whatever you take would exceed the allowed mining angle.
As already explained above, for a human expert it is very obvious what is needed to solve that problem: Do some more wasting ahead of time! So this is a perfect case for adding some "strategic waste" to the strategy.
But before we go for "strategic waste", we still try something else. Looking again at the lower right screenshot we also see that the resulting mining strategy is also very "ugly", not nicely following the existing benches, but in some places move forward, leaving other places behind, generating a very much "rugged" mining front.
For such cases we already had in the past an extra option for the "short term waste" which is "waste to preserve priority". With priority being defined in such a way that always blocks along the mining face should be tried first, before moving further.
The result is indeed much "nicer" - with the drawback of terminating even earlier, after about 2 1/2 periods already, with a total production of 726 ktons at the end.
Now we have added a number of "strategic waste products", and with this it is not a problem to achieve almost 9 periods, with a total production of 2586 ktons and taking out almost all the available material from the quarry. We kept the "waste to preserve priority" option on, and the result is accordingly: Mining looks more or less regular.
Note that also with strategic waste on, we are still using the "short term waste" option on top of it: You can see this in the chart below the sequence of screenshots.
Hint: Click on the image to open an enlarged version in a separate browser window or tab!
Finally we still want to see what happens if we turn off the "waste to preserve priority" option off also here.
The result is a mining strategy that is again a bit less regular, and production is again slightly higher: We are able to finish also the 9th period, with a total production result of 2700 ktons.
Conclusion
This is now the best we could achieve - after a lot of changing waste ratios for the different "strategic waste products"! Still we do not fully achieve the 10 periods that should be possible, theoretically.
The reason is that we allowed to also waste some of the "Mg Limestone", while the "Optimizer" had told us that we should be able to use it all - illustrating again the fact that the "Optimizer" is really important for the strategic planning process! But I was losing my patience - and it is still just a synthetic example - right!?
In real life this would not be a good excuse of course! So in any case I would not allow wasting of "Mg Limestone" in the first place. And then I would not insist on simulating the entire quarry lifetime with one single "Scheduler" run, but rather do it in several steps.
Still I hope that with the above example the concept of the two waste types could be clarified a bit more than just with a theoretical explanation! And as already stated in the beginning: It is not making "wasting" more automatic, but rather adds another complex tool to the toolbox of the planning expert.
