Tag Archives: ore characterisation

Modernising last-generation geo-metallurgy practices

By Wolfgang Baum*

Dan Gleeson wrote in International Mining April 2020, “…near identical flowsheets have remained the status quo for decades, with the only variation tending to be how many pieces of conventional equipment are used, as opposed to what new innovations are slotted in up- or downstream of primary crushing”.

It is puzzling that, in many mining companies (and the EPCM world), the ore characterisation status quo has also remained the same since the 20th century ended. Robust ore profiling via modern quantitative laboratory technology is spotty at best or frequently outsourced to commercial labs. Models continue to be filled with too much visual data and lack sufficient quantitative process-related details. Hand lens and pocket knives or other qualitative tools prevail, while we try to utilise Big Data analytics.

Geo-metallurgy, the most overused word in mining, lingers on without emphasising the heart and essence of geology and metallurgy, ie modern ore characterisation focused on processing and based on instrumental laboratory work.

The current situation in ore characterisation is reminiscent of the US steel industry, post-World War II: the status quo remained with companies bound to familiar technology. When ‘big steel’ removed its blinders, the industry and technology had changed. Geo-metallurgical work, if it is to make meaningful improvements in future mining, requires ‘Ctrl-Alt-Delete’ followed by some seismic optimisations.

The low-hanging fruit has been harvested and future orebodies will not forgive one metallurgical mistake. A paradigm shift is needed toward robust continuous ore profiling (chemical, mineralogical and metallurgical). And, routine ore characterisation has to range from the blast holes and draw points, to the rougher and final tailings.

The few positive exceptions

Of course, a few exceptions stand out and need to be applauded.

Several mining companies have implemented and advanced cutting-edge laboratory technology and lab automation, built 24/7 central laboratories and continue to modernise ore characterisation in the direction of cross-belt analysis, downhole logging and large-scale orebody profiling.

Yet, these are exceptions, not the rule.

  • From 15 new copper concentrators built during the last 12 years, only +/- 14% had mineralogical lab capabilities on site. This contributes to delays in reaching nameplate capacity, more downtime and ‘noisy’ metallurgy; and
  • In 2018, only +/- 50% of the 10 largest copper producers and only +/-25% of the 10 largest gold producers had significant modern laboratory automation and process mineralogy labs at mine sites.

Too often, lab work has been minimised, de-prioritised and/or run as shortcuts by many mining companies.

  • Block models continue to be overloaded with geochemical data and a lack of sufficient quantitative mineralogy;
  • Operators wonder why they did not receive a warning from the mine geology department that the swelling clay content increased by 4%;
  • Most modern haul trucks may have over 200 sensors, yet most mines lack a cutting-edge mineralogy lab;
  • In 1973, Don Hausen pioneered the use of large-scale XRD (X-ray Diffraction) alteration contouring in Arizona. Forty-seven (47!!) years later, feet are still being dragged on installing routine XRD equipment for mine geology and processes;
  • Heap leaching in copper (and some other metals) continues to be challenged by wrong placement of ore, leach test coding errors, over- and under-crushed and over- and under-cured feed, poor agglomeration, high acid consumption or inconsistent acid addition, scattered permeabilities and permeability failures;
  • Many permeability failures in heap leaching were caused by a lack of geological ore control, missing quantitative clay data, ore blending based on visual logging, and/or poor leach practices;
  • Visual diagnostics are the most inadequate tool for identifying and monitoring detrimental minerals such as talc, pyrophyllite, swelling clay, hornblende, zeolites, acid consumers, pH–changing minerals to name a few; and
  • Finding out, post-start-up, that the mine has a poor hardness, pyrite or clay model should not be acceptable in the third decade of the 21st century.

Not modernising last generation laboratories imperils mining from pit to plant, and is one reason for underperforming flowsheets.

Having a few dump trucks or conveyors misrouting ore in large, integrated copper stockpile leach/heap leach/concentrator operations with molybdenum by-products may result in rapid compounded losses in the multimillion-dollar range.

Pushing tonnage (‘tonnage farming’) has its place, but, without concurrent good ore characterisation, it can be a high risk to optimal metallurgy.

For geo-metallurgy, labs are not everything – but they are an extremely important thing.

The plant does not see assay Cu, it sees minerals and textures!!

If modern ore characterisation had been used as part of an integrated geo-metallurgy program at many mine sites, there would have been significantly less of the issues listed below – most of which are related to lack of mineralogy data and continue to drain money from many mines:

  • Excessive frothing;
  • Over-reagentising;
  • Reagent pyramiding;
  • Permeability failures;
  • Undersizing of cleaners;
  • Higher pyrite dilutions;
  • Inconsistent feed size;
  • Runaway float conditions;
  • Self-floaters being out of control;
  • Unexpected changes in PSD;
  • More oxide molybdenum than modelled;
  • Oxide ore sent to concentrator;
  • High clay affecting mill loading;
  • Higher clay content than modelled;
  • Lack of control of acid consumers;
  • Excessive salting in heap leaching;
  • Increased wear through clay in high pressure grinding rolls;
  • Higher self-floaters = increasing smelting cost;
  • Feed rate variance 12-35% = incorrect mill sizing;
  • Higher acidic gangue = xanthates become ineffective;
  • Uncontrolled clay/mica/chlorite = sluggish molybdenum float;
  • Reduced grinding efficiency – packing of clays in lifters;
  • Excessive sliming = problems with cleaner scalper function;
  • Contaminated recycle water – higher O/F turbidity – U/F density issues; and
  • Poor selectivity, brittle froth, gangue entrainment and too much pH variance.

The cost of the above ‘issues’ may be in the hundreds of millions of dollars range.

Fire prevention, (modern lab technology at mine sites) on the other hand, is cost efficient

A heap failure due to clay variance can cost upwards of $15 million, in some cases, whereas unplanned concentrator shutdowns due to de-bottlenecking and tailings losses can prove even more expensive.

Ore misrouting, one of the larger loss factors due to a lack of routine process-related ore characterisation, may increase with deeper pits, more underground operations, longer hauling/conveying and higher strip ratios.

Big Data analytics requires ‘large-spatially-gridded-sampling’ and quantitative mineralogical and chemical characterisation of these samples. Anything short of this, despite the assumed cost savings, remains risky over-simplifications which lead to ‘speculative models’. Efficient geo-metallurgy will start when we enter process-related ore characterisation data into the models.

In regard to cost-cutting of on-site laboratory services, it’s just like in the nautical business – you can’t keep your sails trimmed forever. Future geo-metallurgy efforts would benefit from an assessment of ‘lessons ignored’.

*Wolfgang Baum is Managing Director of Ore & Plant Mineralogy LLC