Dewatering: Thickening, Filtering, CCD, Water Treatment & Tailings Disposal

Dewatering: Thickening, Filtering, CCD, Water Treatment & Tailings Disposal

  • To participate in the 911Metallurgist Forums, be sure to JOINLOGIN
  • Use Add New Topic to ask a New Question/Discussion about Thickening, Filtering or Tailings and Water.
  • 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.

Predicting Future Sulfate Levels (4 replies)

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

Potential Major Sources of Error in Predicting Future Sulfate Levels in neutral mine drainage.Any input for the questions below would be appreciated.Not counting flow estimates and predictions, what are some major sources of error associated with modeling- based predictions of the composition of future mine process /tailings water

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

The first source is sampling itself. How representative are the samples being tested of the mineralogies and textures of the field-scale rocks that will be modeled? Given the realities of mining, this is a central issue. In virtually all circumstances, there will be a relatively high density of drilling and sampling associated with ore zones, but a much lower density associated with waste. This makes perfect sense with respect to establishing the resource model, the sine qua non for a prospective mine. However, the geo-statistical rigor that is required today for a resource model is rarely or never available for waste rock (although almost always waste will be 90+% of the rock managed, certainly for large volume, low-grade open-pit mines.

Then, how has design of the HCT affected the representativeness of the samples? For example, you need to restrict the size of the fragments that are under leach in order to obtain reasonable fluid-flow performance in the tests. Fair enough. Consider a stockwork-veinlet system (for example, ubiquitous in porphyry-type systems and also many epithermal). It often is observed that the sulfides will be concentrated in the veinlets, whereas the phases with acid-neutralization capacity are preferentially located in the matrix. When this occurs, natural breakage often is preferential along the veinlets, exposing sulfides preferentially and isolating the ANC. However, over-crushing of the sample (e.g. using jaw crushers) can overcome this, providing an ANC/AP ratio that is very much biased toward the ANC. The questions associated with how to scale release rates from HCTs for surface area (or even whether to do that) are widely discussed in the literature over the last 10 years or a bit more.

We know that the problem you mention, occlusion of carbonates by secondary reactions, can occur, because it is well known from aerobic limestone drains. One can imagine that there also are circumstances in which secondary reaction products occlude sulfides (not to mention in which either sulfides or carbonates are occluded by other minerals prior to any reactions occurring). There also are issues with how effective occlusion may be when the dominant physical mechanism is diffusion and the long-term stability of the secondary phases in a transient geochemical environment; these has been issues in demonstrating the efficacy of various proposals for “passivating” sulfides. These matters, and many other relevant ones, have been addressed for many years in the geochemical literature on weathering in the critical zone, and the evidence from those studies is well worth studying in terms of understanding how physical and chemical processes interact in weathering of mine wastes. A good introduction is White and Brantley’s 1995 overview chapter in MSA Reviews in Mineralogy, Vol 31, Chemical Weathering Rates of Silicate Minerals. A major finding of studies pioneered by Art White (USGS) is that when calculated by inversion from watershed-scale studies, the implicit rates of reaction are typically 1-3 (or sometimes more) orders of magnitude slower than the laboratory-determined rates for the same minerals. There is evidence for this also in large-scale mine-wastes (although this does NOT mean that mine wastes are not a threat to water quality: the mass-action consequences also need to be considered). Putting aside large-scale hydrogeological uncertainties (as you propose) is fine. But we cannot usefully exclude understanding how the physics of solution-solid systems is related to the geochemical reaction that we propose or infer.

S
Sturmbann
8 years ago
Sturmbann 8 years ago

As always, you provided an insightful response. To expand on it, I think the following thoughts are also worth bearing in mind.

Relative to resource estimation, water quality prediction has the added uncertainty of simulating geochemical reactions, which can be based on rather nebulous thermodynamic constraints. Even with geostatistically 'adequate' data we face additional hurdles compared with measuring total resource content.

Using humidity cells to measure kinetics and then to predict water quality is dependent on a lot of assumptions and verifying them can be onerous. Unless you are dealing with a sample that is both physically and geochemically representative of the reactive material in the field and the leaching environment is sufficiently similar, it is reasonable to expect some differences in leachate composition. Water content, flow rate, temperature, oxygen supply, biological activity, trace element distribution, particle size distribution, etc. can all be important controls, aside from basic mineralogy.

The leachate concentrations from kinetic test methods are also not necessarily measurements of reaction rates, but can also be dependent on solubility constraints. When dealing with water quality, the line can be blurred between these points of departure, as flow rates in the field may be slow enough for reactions that are 'kinetic' in the lab to be viewed as equilibrium reactions in the field. Slow / un-measurable reactions in the lab may also become 'kinetic' under field conditions. The issues of diffusion controlled reactions and armoring also need to be considered in this context, following on from comment above.

It is also worth considering that humidity cells and open field cells are generally oxidized, but that may not be an accurate reflection in a dump for instance, where the Fe2+ / Fe3+ buffer may control the oxidation state at an oxidation front. This is important in your case, as it can be the difference between limestones becoming armored or remaining available for reaction. Differences in water content and, by implication flushing, can also mean that different mineral solubility boundaries may be relevant between field and lab conditions.

In practice I don't think anyone realistically tries to model all these processes in the consulting world, but it is important to consider how sensitive your system is to them and to decide which ones need to be accounted for.

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

Perhaps a more useful framing of your question is, what the sources of uncertainty instead of error are. The projection of geochemical processes into the future is fraught with uncertainty. It is often useful to think about the potential outcomes in terms of a range in which 80% of the potential outcomes could be encompassed. Professionals are pretty poor at estimating this range (they tend to estimate too narrowly) until they receive a bit of training, but once trained and exercising this knowledge in combination with professional judgment in estimating the range, the question to ask is - would I (or the project) do anything differently if I was at the top end of the range as opposed to the bottom end of the range?

If the answer to that question is no, then you are pretty much done. If the answer to that question is yes, then it is often useful to understand which data collection or analysis could reduce that uncertainty, and by how much; and what specific development or operational decisions would the reduction in uncertainty affect. Using Value of Information, you can then estimate whether it is worth the money to collect that additional information, and/or do that additional analysis. If it adds value, then collect the info. If it doesn't then acknowledge the uncertainty and make the decision.

As scientists, we often fall into the trap of thinking that more data and analysis is always good and that uncertainty is always bad. In practice, it is often better to embrace and understand the implications of that uncertainty to decision-making, and to help the project move along when collecting additional data doesn't have a good chance of changing outcomes.

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

Thanks for elaborating important issues. One quick thought on your important comments on understanding the conceptual basis of kinetic tests. It is of central importance that design and excite test work with some specific intentions, and that those intentions are consistent with the thermodynamic and kinetic processes that are actually likely to exist

In the test and

In the field-scale facility that we actually need to understand.

One way to consider the HCT environment is that what we analyze form the effluent (flow and chemistry) is a *release* rate, not a mineral-specific dissolution rate. Whether the secondary constraints that are operative in the HCT also are operative in a waste-rock or tailing pile is critically important to understanding how to apply the test results to the system for which decisions need to be made.

I also am a believer in Value of Information, and am very pleased that you brought it up. Too often we see the geochemical testing and analysis (including modeling) to be an abstract activity that seeks to uncover a "Truth." This is a very heavy burden. At least in supporting the planning, operation, and closure of mines, usually what we are trying to do is to support decision-making: do I need to isolate material X with major engineering controls, may I manage it with discretion, or possibly, may it have specific properties that imply positive values for another aspect of integrated waste management? Within each of these (and other possible) classes of decision-making, we usually have a range of possible characteristics that would be consistent with the proposed end-point. We need to be *accurate* with respect to how we categorize the material, but we may not need great *precision* within the category. As I understand it, is that data and information are not synonyms. What we need is the cut-set of data that allows us to formulate the analyzed information needed to make good decisions at a known (or at least estimated) level of reliability.

Far from being tangential, your thoughts are central.

Please join and login to participate and leave a comment.