- To participate in the 911Metallurgist Forums, be sure to JOIN & LOGIN
- Use Add New Topic to ask a New Question/Discussion about Flotation.
- 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.
Flotation Regrind Circuit Cyclones (13 replies)
A good place to start might be firing an email off to the hydrocyclone manufacturers. Krebs, Warman, Linatex, etc. They probably do more test work than most.
You must have meant P100 less than 300 microns applicable in primary grinding circuits, not UFG circuits. In a recent UFG application, we had about 25-30% solids in the cyclone feed and 9% solids and 12 microns in the cyclone overflow. Operators usually have quite a lot of tools and flexibility in their circuits for optimization. However their beliefs (i.e. personal choice, past experiences) can determine how the cyclones should be operated and this can become obstacle for them to go outside of their comfort zones to find out best operating conditions, applicable for their own specific plants/circuits.
Perhaps also talking to anyone making Mozley or Salter style desliming cyclones (think 2 inch body, UF split). I had found the Salter guys in the UK pretty helpful when using their product. I can't seem to find any of the info I had back then though.
I had quite a bit of data from MIM and Century in the past but have lost it. From what I remember, yes, the cyclone feed % solids are significantly lower in regrind circuits, say -30%. The underflow % solids were typically about 60%. Mass splits from what I remember were typically about a third to overflow and two thirds to underflow. The regrind cyclone overflow, which is based on the concentrate % solids feeding regrind circuit including launder water etc is relatively dilute - 20 - 25% solids or less esp in fine regrind circuits. As a result, the water split to underflow is quite a bit lower in regrind duty (-< 15%) compared with typical primary grinding duties (-30%). Hence lower bypass to underflow and an overall sharper in corrected efficiency curve. You might want to try Dr Bill Johnson at Mineralurgy in Brisbane, I'm sure he would have some data. Hope this helps.
When I said P100 what I meant was 100% passing in the regrind cyclone feed. I probably should have said F100 for clarity. I know this is still a large range but typically from the data I've seen from large rougher cells (mostly copper) you do get some +150 and even +212 micron particles reporting to concentrated hence I'm assuming that 300 microns is the top size going to the regrind cyclone even when the F80 may be significantly lower.
Encourage you to engage US based metallurgist Metcom technologies who specialises grinding circuit optimisation and training. They have some excellent papers on the subject.
The cyclone does not know whether it is processing/separating grinding circuit product or a flotation product.The feed characteristics/operating and design parameters decide the performance. Barry Wills Text book describes the characteristics of cyclones and a reference to such a book is suggested.
To manipulate the solids and water split one would have to change the cyclone cut point (D50) and the cyclone spigot (apex) diameter. To change the D50, we have the flexibility in our cyclones to change configuration, i.e. diameter, cone angle, vortex finder diameter etc. We use software that has been developed for our equipment to do first order cyclone performance predictions, based on these results we can then make a recommendation which is then further optimised in the field (if required).
For fine cut points, one would require / employ small diameter cyclones. Problem with this is that one would have to employ a large number of these cyclones, in parallel, to obtain volumetric capacity - that is where cluster cyclones come in.
One issue that you will always see in flotation is that the water added to the flotation launders in the different stages is not measured and sometimes you can find that flotation cells and cyclones are working out of their range of operations Under controlled conditions, the cyclones should be able to provide you with a predictable performance
The operators are smart people with limited time to learn and to analyze data. They will not be able to help you to have the cyclones working properly.
So make sure that you know the process variability and that extensive sampling is done in the plant to determine if the cyclones are properly configured to perform as per the input parameters at which the slurry is fed. This may be the largest issue and not the model. I look forward to receive data to attempt using my models to fit the actual cyclone performance.
You make an excellent point with respect to launder water addition. I'd even go a step further and make the point that if the launders in the float cell (and launders outside of the cell) are designed properly (with sufficient slope) that they don't need additional water to keep conc moving. There are some cases where the froth is particularly strong or sticky and needs some water to break them down but in my experience these are few and far between. In many of the plants I've seen the water is needed for the pump, either to get the suction head right or to get the flow up to a volume the pump is comfortable with. This brings us to the debate on plant design which is almost a chicken and egg scenario - is it better to design the plant with a smaller size and lower flows or do we overdesign and put in a bigger pump knowing that we can always add water. It’s an interesting engineering problem, made even more interesting when considering that for many plants today water is scarce > maybe we (as an industry) need a design paradigm shift? Anyway I'm getting a little off topic.
No need of the reminder of who you are or the work you've done > I'd even go so far to suggest to any young Minerals Processors read up on the cyclone work from you and Lynch. Since we have the benefit of your experience in this discussion I'll pose another question; My understanding is that one of your findings was feed size had a major effect on performance (not surprisingly) with the empirical model you and Lynch derived two of the fitted parameters were for constants multiplied by a % of material in certain size classes (+420microns and -53 microns to my understanding). Did the work ever look at experimental data where the +420 fraction was zero and the -53 micron fraction was close to 100%?’ In other wordsheading toward the regrind scenarios we encounter today.
Thanks for your remarks; To be honest, both of us moved on to other areas and we still feel(we are in constant touch) that basic contours are there in our model and all one needs to do is conduct a test on the size distribution he has to treat and then the rest becomes obvious. Many cyclone manufacturers follow our model as a base and using their data bank, they developed their own curves to select cyclones.
Innovative technology to reduce over grinding:
After 20 years of experience in Flotation and grinding, we were given a challenging task by both Govt of INDIA and HZL (VEDANTA NOW). How to reduce silica in zinc concentrate, as smelter rejected 8000 tons of concentrate, due to high silica in concentrate ISM > 8%.
We could solve the problems through COLUMN FLOTATION.
We further did R&D commercial for RGBM circuit to avoid generation of ultra-fines. The Most dangerous particles for flotation process!
INNOVATIVE CKT: RGBM cyclone was converted as de watering unit. Under flow was fed to Zinc Rougher cells, (In old it was fed back to Ball mill), Over flow having ultra-fines with more water was fed to SCAVENGER cells. This gave surprising results 50% Zn, improved to 53% and Tailing loss from 0.5 to 0.3% Zn. Anyone can try this and find advantages. You can do even LAB tests and confirm results, and compare with closed circuit.
The problem is that these units run under very different conditions to most primary cyclones, generally with lower % solids feed and a much tighter size range (P100 less than 300 microns). What we were interested in is predicting water and solids splits from these units in conjunction with fine and ultrafine grinding found in flotation regrinds. It seems that many of the existing models appear to have been developed for primary grinding applications and over or under predict these splits. Can anyone point us in the direction of any recent work done on cyclones in this type of duty?