- To participate in the 911Metallurgist Forums, be sure to JOIN & LOGIN
- 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.
Statically Valid Method to Verify a Primary Sample Cut (11 replies)
Have a look on the sampler, how it is designed and working, how it is cutting the stream. It is the first step to highlight eventual dysfunctions, mistakes in design and operating conditions. If the sampling strategy can easily be configured, you can use the sampler itself in various operating conditions to perform a bias test.
The bias can also come from mine side. To check that, you have to verify if there is a bias between mill feed and final products (concentrates and tailings). In that case, data reconciliation by material balance can be required.
He has given you the right advice. The main observation is to see if the primary sampler captures all the size fractions from the stream - what type of cutter is it? Cross stream or belt?
As a geologist my suspicions regarding bias usually go towards the mine side due to the difficulty of getting representative samples, particularly in underground and when using blastholes. The key questions are what type of sampling and estimation methods are being used there to determine the head grade? What material is not being captured in the samples in the mine that are used to estimate the head grade?
If reconciliation shows important differences, you must analyze the two processes being compared:
Plant Sampler: You should check the sample from the primary cutting to mechanical preparation. In primary cutter, you have to see if the opening captures 3 times the largest particle diameter (D95), if the cutter is linear and completely cuts off the flow, if the volume of the cutter is sufficient to retain flow cut off, making sample is systematically over time. Vezin should be reviewed.
Mina: You have to see, the technique of sampling and QAQC, estimators, dilution and lost. Survey is conducted daily on the progress of extraction.
We must analyze all the processes to determine which of those affected.
Mine to final product reconciliation is a long process. Implementation of a Metal Accounting system is a way to better estimate the mine and plant performances as well as the stocks and WIP. As it is based on various types of measurements (weighing, volume, moisture content, assays), such an implementation is really efficient if the system of measurement is audited to check all sources of errors and to quantify these errors. Due to the current metal price, this kind of audit has a very short ROI.
Thank you for your feedback. Paid metal would suggest that we have a low bias. We currently use our material balance to adjust; however, due to downstream processes we are not able to run a daily balance to filtered concentrate. We are installing additional instrumentation and samplers which will allow us to do so in the future.
The plant feed samplers are crosscut design followed by two stages of vezins. The samplers were designed based on feedback from Pitard and Bongarcon after audits of the system we previously had.
We have conducted extensive audits of our sampling system from collection through assay. The bias does not show up as a function of the steps along the way leading us to believe it is in the primary cut.
One thing you might look at sampling is initial sample size and methods you are using to acquire sample, and methods used to reduce sample to size you can carry to lab. One of the big problems taking samples from blast area from a pit can be; if you sample from a pile larger heavier particles will fall to rest near bottom of a pile. One of the ways to mitigate this is have samplers place a plate in the pile to keep debris from falling from above and digging into the pile for your sample. A good source for determining what size sample and how to sample are the ASTM methods for aggregates. http://is.gd/2PzfZ5 (ASTM D75/AASHTO T 248). While specifically for aggregate samples, these methods are applicable for taking any sample where segregation can occur.
If Francis Pitard and Dominique François-Bongarçon were involved in sampling system designing, you can be confident. You were in good hands. The issue can come from another point: mine? Slurry sampling!
Sounds to me like your problem lies at the drilling stage. Are you sampling from blast holes? If so, look into RC Grade Control. Have a look at http://is.gd/Xux8qG. What we do know with any shadow of doubt is that blast hole samples are completely unrepresentative no matter what sampling equipment is in place thereafter, i.e. what is coming out of the ground is unrepresentative (please discuss this area with Pitard and he will agree). Therefore, no matter what you do with this material thereafter to obtain data about your orebody, it will be wrong along with every subsequent decision. Remember, even if it were practical to collect and assay the entire bulk sample from every blast hole, the results will still be unrepresentative. The reasons are many and are explained in the presentation.
I may be able to assist with information related to real time analysis of the full stream of primary material on the belt see http://is.gd/XvZrPn
Do your ore mark-outs adequately take into account blast movement of ore using blast movement balls? http://is.gd/ASBtnF If not this could result in a systematic underestimation of contained product being feed to the primary sampler.
One clue I have used in the past is to look at the mineralogy/material types of the sample vs. the mine. This can often do a better job of indicating whether the sampling is introducing a bias - e.g. sticky/fine material types hang up in the cutter or have a different trajectory which does not get into the cut (or too much in the final sample). In other words, get a geology version (not just the assay!).
I also agree that in-pit sampling - whatever you use - is impossible at best (says the metallurgist)!
We have used interweave sampling to verify validity of secondary and tertiary cutters with regards to the primary sampler, but still have a consistent bias in assays between mine and mill. This still leaves the possibility of a bias coming from the primary sampler. Has anyone done any work to try to verify accuracy of primary samplers knowing that it is impossible to capture the entire feed stream? (e.g. verify that the primary sampler is representative of the entire stream being sampled.)