Geology & GeoMetallurgy

Geology & GeoMetallurgy

  • To participate in the 911Metallurgist Forums, be sure to JOINLOGIN
  • Use Add New Topic to ask a New Question/Discussion about Geology and Geometallurgy.
  • OR Select a Topic that Interests you.
  • Use Add Reply = to Reply/Participate in a Topic/Discussion (most frequent).
    Using Add Reply allows you to Attach Images or PDF files and provide a more complete input.
  • Use Add Comment = to comment on someone else’s Reply in an already active Topic/Discussion.

Reconciliation uncertainties in mine production (12 replies)

Raje Singh
8 years ago
Raje Singh 8 years ago

I am working on a Reconciliation uncertainty associated with mine production and I want to know where technical and geological activities come into play.

O
OberstGruppen
8 years ago
OberstGruppen 8 years ago

What aspects are you trying to reconcile - ore tonnage, tonnage of specific mined areas, grade, etc.?

(unknown)
8 years ago
(unknown) 8 years ago

Are you trying to reconcile production from an open cut mine or underground or from multiple sources processed as a blend? Reconcile grade control estimates with mine production, with stockpiles and processing. Need to be specific as many factors apply.

A major weakness for any reconciliation from an open cut is you do not know how much ore has been directed to a waste heap. Any equations generated are poorly constrained and only as good as your assumptions. Do you use Blast Monitoring Balls to assess the blast movement? How are you mining it? Do you mine on flitches or benches and do you make adjustments for blast heave in an open cut, otherwise you will likely over dig the upper flitch.

A major weakness for an underground operation is you may not know how much unplanned waste dilution has been added unless the cavity volume has been determined by a cavity monitor. Reconciling from multiple blended sources is a major challenge. PS being a geologist you will be blamed for any adverse results no matter what.

Oberfuhrer
8 years ago
Oberfuhrer 8 years ago

Basically the uncertainties arise from incorrect or poor data capture, and little awareness of the range of values and how small differences in inputs can lead to large errors.

Essentially, SISO: Sh*t in, Sh*t Out.

For example, there is always uncertainty around the mine model (it is a model after all) based on the potential errors in drilling, sampling and interpretation, then the losses during mining, which may not be captured well - mis-dumps mean less of the resource through the plant.

Along with this, the sampling of tailings and impact on recovery calculations can be significant; if product reports to tailings in an amount different to the calculations that will have an obvious impact.

(unknown)
8 years ago
(unknown) 8 years ago

You nailed a lot of stuff to consider. I would also add the techies have a habit of picking the shiny when sampling stockpiles. If your ore has shiny minerals (pyrite, galena etc) you will find that people will naturally pick those rocks out. Make sure your samplers are taking your stockpile samples correctly and with no bias.

Ps. Size of the rocks can also make a difference if the host rock is more competent that your orebody as the May only take rocks that fit in the bags and not take a true sample of the stockpile.

Watch your samplers closely.

Maya Rothman
8 years ago
Maya Rothman 8 years ago

What is your drilling/sampling method for grade control, i.e. Blast hole cone grab sample, RC GC with cyclone/riffle, etc? And how far out (%) is the reconciliation between geology vs. mill? Is this a gold application?

U
Unterstarm
8 years ago
Unterstarm 8 years ago

Beware of making short term reconciliations, as it may not give you the correct picture. Management can be quick to jump on negative reconciliations, so it is your job to educate them that reconciliation has to be looked at over a longer period of time. The issue can be complicated if you are feeding from stockpiles, mining is taking place from one small area that has grade and material variations not reflected in the broader model etc.

Agree with you about the SISO issue. Make sure you have the correct systems in place to capture and analyse your data.

Treatment of 'batches' through the plant may also be helpful - but make sure the batches are of sufficient tonnages. I would suggest at least several days of material.

(unknown)
8 years ago
(unknown) 8 years ago

If operating from stockplies near the ROM dump pocket it is best to establish finger stockpiles and know what was mined and stockpiled so that it can be tracked back to where it was mined in the pit. The stockpile should be closed from adding ore and surveyed before ore is removed and processed. While the stockpile is being depleted and processed another finger stockpile can be built. Make sure that each finger stockpile is separated and of a regular shape for ease of surveying at closing off stockpile or at the end of the month. This will facilitate treatment of batches as earlier indicated. Tonnes mined and trucked the trucking factors can be constrained by the pit survey pickup.

Engineers can give you many challenges by part direct tipping ore and drawing from poorly constructed stockpiles. Reconciliation is only as good as the effort to track each package of ore from mining, stockpiling and processing. Last on first off depletion of stockpiles will generally give poor reconciliations compared to opening and closing off stockpiles.

Mining open cut ore the success of reconciliation will depend on how well you can constrain mining ore and minimise dilution and ore loss. It is important if mining a bench in two flitches the blast heave is appropriately accounted for to prevent over digging the top flitch. Use of blast monitoring balls to account for blast movement either by translation or morphing the blast movement will minimise loss of ore to the waste heap. As pointed out previously reconciliations do not take into account the amount of ore directed to the waste heap.

The secret to a good reconciliation is how well you are prepared to track and manage each stage in the movement of the ore. Getting cooperation can sometimes be difficult as mining KPI is about BCMs and tonnes and processing is about recovery. The geologists are in the middle.

Raje Singh
8 years ago
Raje Singh 8 years ago

I am actually looking at ore tonnage reconciliation-- grade control estimates versus mine production and processing. This is a gold production business and we are operating an open cut here, and we do carry out RC drilling campaigns in our operations and I do agree perfectly with you about a whole lot of errors bound to happen from drilling to even mining, but we are a lot more critical on that and I think we are on top of it though there could be minor errors. Our sequence of mining here is the heave + 2m representing our first flitch and the rest of the 4m for our second flitch in a 6m bench. Could there a better alternative??

For our blast monitoring we make use of monitoring pipes but I think I agree with the best option would be blast monitoring balls as the monitoring pipes could be bending. We would work on that. The blasting itself has been a challenge too, because poor fragmentation sometimes occurs and that leaves some toes on the ground. We have come up with a QAQC system that checks hole depth, charging height and stemming height. Are there any systems that can improve fragmentation?

Zander Barcalow
8 years ago
Zander Barcalow 8 years ago

Your comment about short term reconciliation is very important in the mine to mill process as well.

If your residence time is more than a day or so from crusher to tails then the amount of material not "in common" can mount up pretty quickly. At Boddington the process residence time is about two days, so in a month two days of mill reporting is based on material from the previous month and two days of crusher feed at the end isn't reported fully in the mill figures. This means that for a 30 day month 13% of the material is unmatched in the reconciliation. To combat this we focus more on the three month rolling mine to mill trends as the major trigger from a reconciliation perspective, with only major variances in the shorter (daily to monthly) measures seen as a cause of concern (for example a major feed grade change not appearing in head grades over a week.)

Oberfuhrer
8 years ago
Oberfuhrer 8 years ago

You’ve been lucky to get some top notch comments; they have been dead-on, residence time in processing and stockpiling strategy can make it next to impossible over the short-term - over the long term, both your wins and losses can be smeared out.

With ore tonnage, are you comparing grade and metal? For example, if you mine/process more but get the same ounces, you'd take it - opportunity to improve mining hygiene etc to lower process cost, but at least the metal is going to the plant rather than the dump.

Zander Barcalow
8 years ago
Zander Barcalow 8 years ago

The tonnage reconciliation issue is always a challenge to overcome. Getting your truck factors right is often difficult as systems such as VIMS work on a wet tonnage while your model report/truck factors will generally be a dry tonnage and the difference can vary with the weather. If you do use VIMS always use a large data population and check its descriptive statistics as it is normally a normal population with a large variance, and check each truck to ensure they are measuring properly as 1 or 2 bad truck weight-o-meters can really influence the result.

Have you conducted any check on the swollen volume of the blast to calibrate the flitch tonnages? Use the difference between the pre-blast and post blast volumes to calculate a swell factor and have a look at whether your flitch RLs are correct. Assuming a 45% swell on your blasted material it looks like your bench RLs are fairly close accounting for slightly less swell in the second bench. Creating the volumes from the survey surfaces is difficult as the blast will tend to ramp over the shot in front unless you freeface the blast (and spread it halfway across the pit floor.) Alternatively using BMMs installed at your midbench you could directly measure the vertical movement for setting this.

A second factor may be the matching of load unit fleets to trucks. Looking at your VIMs data, if available, could help identify variances between load unit/truck fleet matching that may be influencing your reconciliation. Even the individual bucket sizes within a fleet could have an impact.

For fragmentation and blasting toe they are probably focussing on the blasting issues you are facing. We have had a vast improvement by intelligent trial and error, changing everything from stemming height and blasthole spacing to firing timings and hole initiation patterns and measuring the results. Unfortunately balancing fragmentation and movement/dilution is difficult and the outcomes will need to be tailored to the economics of your deposit/operation.

If you have got significant density differences across your deposit then the blast movement could also be influencing your reconciliation as, even with modern BMM systems, most of the time you will reconcile tonnages against an un-translated block model. This will mean there are significant edge effects where the material you mine does not match the material being measured. For example the translated blocks move 5m forward and are mined out and the survey face used for reconciliation does not cut the relevant blocks in the block model. This should be rectified in longer term measures (3 or 6 month rolling vs. 1 month) if it is occurring and should be explainable to management.

(unknown)
8 years ago
(unknown) 8 years ago

The general opinion is that long term reconciliation is more reliable than short term reconciliation, but compensating errors can hide a multitude of sins too. One needs to be careful of reacting to short term reconciliation and getting into analysis paralysis and taking your eye off the geology. As indicated above management will react to a negative reconciliation. The process maybe in statistical process control and any changes may just make it worse as it could become out of statistical process control.

The gold in circuit generally determined at the end of each month should take into account all the gold in circuit, including transitional crushed ore stockpiles, rather than assume as constant. The gold in circuit estimate may not be as robust as metallurgists would have us believe. It is difficult to account for gold in gravity traps if the plant does not have a good gravity circuit such as Knelson concentrators or the like for the earliest removal of gold before it is caught in traps. Sampling gold loading on carbon is difficult as a dynamic system is being sampled and TOS does not apply.

Using black polypipe as a monitor of blast movement can give a false sense of security as it reflects near surface movement which is minimal compared to the middle of the bench, better measured by BMBs.

you mentioned blast fragmentation was a challenge. There are blast pattern optimisation programmes available to minimise mixing of ore, but they all can only throw ore in one direction. However the understanding of the structure and distribution of mineralisation should be input into each blast design. Need good communication between geologists and blast control engineers. Too often each hides in their own silo. If there is cross cutting mineralisation the movement of that ore needs to be accounted to prevent ore loss and ore dilution. This is probably the area of greatest uncertainty assuming the grade control drilling was performed to minimise sampling errors as mentioned earlier.

Truck factors can be adjusted matching total truck counts to BCMs surveyed for the month and expressed on a dry tonnage basis per truck. Samples can be collected and measured for determination of in-situ bulk density. Reconciliation is only as good as your understanding of the geology and tracking each stage of movement of ore.

Please join and login to participate and leave a comment.