Mining Technology reached out to Xore’s CEO Mikael Normark again to find out more about the financial benefits of online analysis. How long time will it take before the investment pays off?
How big can the improvement be with online analysis?
Naturally, it will depend a bit on the ore and the process, for a general case, we typically state 2% better recovery. But we have seen better improvements than that. An interesting point here is that smaller operations can make just as big improvements than the really big operations, and sometimes even better.
The background is that a smaller flotation circuit is more susceptible to rapid changes compared to large circuits. If for instance the head grade changes, a large flotation circuit will take a longer time to adapt to it, giving the operators more time to notice and react to it. A smaller operation will change more quickly and risk running sub-optimal until the new conditions are noticed, and the process is adjusted.
For a large-scale operation, there might be a bit more time to act, but on the other hand, even a small error can quickly lead to big losses. For them, the accuracy is really important, in all the steps from primary sampler to the analyser itself. However, they will also need short analysis times as they often have 20+ sample streams and the analyser should cover the whole lot in 10 minutes at most – without compromising on the quality of the measurements.
How long time does it take for results to start improving?
In a short perspective, there are things that will improve as soon as the first day. One example is to monitor value metals in the tailings. We had a client that during the first night shift with the analyser noticed that the Zinc concentration in the tailings started to climb. After 15 minutes, they decided to make some adjustments to the process, and another 15 minutes later, the zinc concentration was back to its normal level.
They explained that with manual samples they would have needed at least three hours to analyse a sample, because of drying times and handling times. Had they decided to adjust the process and grab a new sample it would have taken them another three hours to get the answer if the adjustment worked or not. So, our analyser effectively made their observe-orient-decide-act loop go from six hours to 30 minutes in this case.
In the bigger perspective, clients start to notice how different ores behave in different ways, and they can adapt the processing. If they are mining more than one lens, they will see in the data from the analyser when a new ore has entered the plant. The metallurgists can then adapt the processing to how it was set the last time that same ore was in the plant. Transition periods can be tricky and without a close eye on the process, it can easily become sub-optimised for a period. Online analysis gives a big advantage here.
For the more advanced sites, online analysis also enables automation based on metal content in the ore. It might be controlling flotation columns based on metal content. It can also be calculating mass balance and mass pull on live data, to decide how much ore to feed to the mills.
How soon can the investment pay off?
Already in the 1990s, there were papers published on this topic. One from 1991 (Jones et al.) says that the typical payback time is between 4.8 and 5.3 months, it is still quoted in current literature. A large-scale operation might have a return of investment as short as 1-2 months, and we see that even smaller operation of, say, 500,000 tonnes per year have a return of investment within a year. So even if times are different now compared to then, I think the figures are still relevant. I would even say that since head grades have generally decreased since the 1990s and mining is going on in even more remote locations, it is probably even more important – and profitable – to take advantage of the technology.
Consider the following example on a plant milling one million tonnes per year, producing 30 kilotonnes (kt) zinc concentrate, 2kt copper concentrate, 2,000oz gold and 400,000oz silver. With today’s metal prices (13 May 2020) and typical deductions for treatment costs, the value of each day’s production is about USD 120.000 for the mine. With a 2% increase in productivity, they will earn USD 2.200 more – every day.
With the investment of an analyser and some auxiliary equipment, sample pumps and piping, a client like this would reach their return of investment in about 9 months.
How can you know if it is a good investment for you?
We can calculate it relatively easily based on the volumes of incoming ore and concentrates, as in the example above. If a client gives us this information, we will return to them with a spreadsheet showing exactly how we calculated it. They can of course also put their own numbers in there. As said before the result is usually in the 2-8-month bracket.
How can a client know if your analyser is suitable for them?
If they provide us with a description of the ore and typical grades in the intended samples, we can say what they can expect in terms of accuracy and detection limits. We can calculate it or refer to ore samples that we have.
I would also like to point out that anyone interested can send in dry samples, we only need maybe 10 grams per sample. We can then make a wax bead, or a couple with varying percentages of solid material, and return with spectrum raw data and analysis results.
The incoming ore is most interesting to us, but you may send in samples of tailings or other materials as well.
We have also had clients send in samples of leaching solutions, process water and electrolyte, so the same procedure is just as relevant for these applications. We would like between 3dl and 5dl of each sample to analyse.
Anyone interested in sending in samples can contact Xore for our shipping address and contact details.