Grinding & Classification Circuits

Grinding & Classification Circuits

  • To participate in the 911Metallurgist Forums, be sure to JOINLOGIN
  • Use Add New Topic to ask a New Question/Discussion about Grinding.
  • 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.

SAG Mill for Small Operation (26 replies)

Helena Russell
8 years ago
Helena Russell 8 years ago

Is it advisable to use a SAG mill for small, say 2000 TPD concentrator operations? 

The processing, in this case, is a complex zinc/tin mineral and we are afraid a SAG would P80 produces too much fines (since the tin is to soft brittles) (tin cassiterite).

The mineral is marmatite and casiteritre, first flotation of zinc, then gravimetry for tin, so we need to avoid the fines even if flotation of zinc needs -100 microns,

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

It's possible, I just don't know how efficient it would be. We had a pilot plant unit that was about 1 m x 2 m and it drew about 80 kW. It worked. The tare power was a little high though compared to grinding power. We used it with run of crusher ore (-175 mm) and it produced -4mm product in closed circuit after screening. It only ran about 1.5 t/h in taconite but produced usable data for upsizing to commercial size.

Short answer, it depends on how much power you have to put into the ore and what the competing process (crushing followed by ball mill, pebble mill, roll press?) is. It's an economic decision. 83 t/h (2000 t/d) is certainly within the realm of realistically sized SAG mills for the appropriate ore.

Maya Rothman
8 years ago
Maya Rothman 8 years ago

For your tonnage I could not see any major benefit. SAG mills are going to be very sensitive to ore hardness and over-grinding is a very real issue. the cost of a skid mounted crushing machine out of china is now so cheap why would you bother. Bang in one of these from China, get yourself a decent stockpile and then a single stage ball mill which will be a lot more stable to run.

We also know that it is better to have too much capacity than not enough.

Single stage mills are proving to be very useful in recent times, as operators are leaning what these machines can do as grind and throughput targets are being achieved.

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

It depends, as always, on the ore. To answer your question in general terms, yes you can use a SAG mill for a 2000 tpd operation, but it may not be the best answer. In this case there may be a need to balance the grinding requirements for good tin recovery vs the requirements for good zinc recovery. Will the ore be treated in a gravity or flotation circuit for tin recovery? Maya provides a good economic approach if the ore's characteristics allow. If not, you can look at SAG using a high ball charge to minimise fines production or even a rod mill/ball mill circuit if you can recover a reasonable quantity of the tin at a coarse grind.

Bob Mathias
8 years ago
Bob Mathias 8 years ago

A straight answer to your question is yes but a definitive answer will require several other considerations.

Examples:

  • ore hardness and its known variability or lack thereof,
  • ore feed size and product size required,
  • possible future expansion beyond the 2000 TPD
  • simplicity of the grinding circuit required( or the suitability of operating only one major equipment besides the classification systems)

These factors when carefully considered can make a suitably sized Single stage SAG mill a viable option.
It has been said by some that the SAG mill is the most forgiven piece of mineral processing equipment that one can have. Once installed it can be readily adapted to meet changing ore and process requirements.

(unknown)
8 years ago
(unknown) 8 years ago
1 like by Muhambet

This was a size that used to be handled by rod mills followed by ball mills just to prevent over-grinding.

Why not look at HPGR to avoid fines?

(unknown)
8 years ago
(unknown) 8 years ago
1 like by Muhambet

The main reason for SAG mills is to eliminate the secondary/tertiary crushing and screening circuits - apparently to save money in both capital and operating costs. May be so in an expensive western country, but unlikely in other parts of the world.

I don't think the throughput is that is critical...I have worked on a 60tph gold operation with a variable speed SAG mill. It was fine.

However what bothers me is the nature of the application - you mentioned tin being present in the feed. If you are interested in recovering a meaningful quantity of this brittle mineral, I would be avoiding a SAG mill like the plague.

Gentle, sequential size reduction and classification - a number of comminution stages with small reduction ratios would be the preferred comminution flowsheet...dare I say three stage crush (or if the project has lots of spare capital a HPGR), then a rod mill, screw classifiers, some jigs, etc, to get that cassiterite out before perhaps a tower mill to more finely grind a selected streams (jig tail?) and recover the sphalerite.

Helena Russell
8 years ago
Helena Russell 8 years ago

Great comments, however it looks like Bill is propossing to recover the tin first, before the flotation of zinc? 

In fact it is a trade of between the recovery of zinc by flotation, first stage, which needs fine grinding to 100 mesh, and the tin recovery (from non float) who requieres not fine grinding, so SAG milling is not a good idea because the tin is a brittle mineral.  It might be a good idea, to avoid slimes (tin fines to much, so to recovered by tables (Deister tables for slimes are excellent), may be with Knelson concentrators of part of the very fine tin and tables for up grading the knelson up graded material.

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

I think that you have confirmed my thoughts...remove the coarse minerals first, with a jig I would suggest rather than a Knelson concentrator.

The mineralogy may allow the sphalerite to be removed with the cassiterite due to mineralogical associations, which would mean a smaller processing plant.

It depends upon what other heavy minerals are present...pyrite/pyrrhotite, magnetite, etc. on what the flowsheet would look like.

The thing about sphalerite is that you need to activate with copper sulphate before it will float (assuming no pre-activation) and thus you have flowsheet and separation choices...sphalerite recovery can be the last processing step. 

Helena Russell
8 years ago
Helena Russell 8 years ago

We had had the comment from Gekko Jig company that starting with jig operation for such a complex mineral (marmatie, sphalerite , cassiterite , pirrothine etc) makes not advisable the Jigging with gekko, however we would test since we had the gekko pilot plant here (sink and float), however I would also consider the best action is start with gross separation, so we at least try, so the avoid fines as much is possible for later recovery of the cassiterite. Also the flotation of zinc as marmatite is rather more difficult than sphalerite, since the concentrate would be less than < 50% Zn. Is there another jig that could be more convenient than gekko for this specific operation as you suggested?

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

All gravity devices will recover the heavy minerals, since the initial cut would be at around SG 3.

This would make a sulphide rich concentrate which you may upgrade, after a coarse grind in a rod mill or a ball mill if a finer grind is required to separate the cassiterite from the sulphides, using a cleaner jig with the ragging set at a SG of 5.8 to 6 or a table.

It depends however on the particle size distribution and this is where the problem of a good separation at a high recovery may lie.

Some testwork is required to see if these are sensible suggestions.

Marmatite flotation is a little harder, but it does float (more copper sulphate required, slower floating and more sensitive to water temperature) - I noted that a Chinese researcher recommended using 'ammonial' copper sulfate at pH 9 (Xiong Tong, Shaoxian Song, Jian He, Feng Rao, Alejandro Lopez-Valdivieso, Minerals Engineering - MINER ENG 01/2007; 20(3):259-263) - the final grade depends how much iron and manganese is present but 50% Zn concentrate should be possible. The sample these guys used had 20% iron in the marmatite, which very high.

Helena Russell
8 years ago
Helena Russell 8 years ago

The zinc/tin mineral that we requested the confirmation from Gekko regarding the gekko jigs application is a mix of complex oxides and sulfides, the cassiterite has a sg 6.8-7.1 and the marmatite (ZnFe)S a kind of sphalerite has sg. 3.9-4.2 , also there is some pirrotine (FeS) with sg. 4.61 and pirite (FeS2) sg 4.95 - 5.10. The ganga is silice (Si02) sg 2.6. 

The general composition of the mineral is aprox. Sphalerite 18%, pyrite 34%, pirrotine 15.60%, cassiterite 2.50% the diference mainly silica. The mineral to be concentrate has maximum 2 mm. With this information Gekko mentioned that would not be advisable to use Gekko Jig. Would be another Jigg?

Regarding the dense separation using cyclone should have the same problems?

JohnnyD
8 years ago
JohnnyD 8 years ago

The SAG Mill use is not the problem itself but the ore complexity and the suitable processing scheme.
In DR-Congo, Kipushi Concentrator has one 18'diam X 6'L Hardinge Cascade-Mill running in the range of 1500-2000 tpd as FAG Mill (design not allowing the use of balls) assisted by: - one Ball Mill before 1992 to treat Cu-Zn sulphide ore by differential flotation; - and 2 Ball Mills from 1998 to make Cu-Co oxide ore flotation.

I agree that SAG can be used for your application because you can not avoid fine or slimes generation having -100 um as milling size target. Combination of ROD Mill and Ball Mill cannot do that.

The complexity of the ore will require sequential size reduction. This has to be accompanied by sequential separation steps. Depending on the cassiterite grains coarseness, you can make: coarse SAG milling and classification; coarse particles submitted to gravity separation Jig-dms'cyclone-spiral-table; regrinding of heavy concentrate if required and differential flotation Zn and Fe and obtaining tail will be the cassiterite concentrate; ball milling of classification fine particles at -100 um followed by Zn flotation and cassiterite fine particles gravity separation.

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

Regarding the separation of the denser minerals what about using a dense media cyclone. at 2000 TPD this should be relatively cheap? That is a good call as a primary separator since SG would be limited to about 3 or so (magnetite or ferro-silicon).

So there a lot of good ideas to work with.

Helena Russell
8 years ago
Helena Russell 8 years ago

I'm not sure what the top size is with dense medium separation, but 2mm should be OK. the problem Gekko refered to in Jigging is that 2mm is very close to the size of the "ragging" or shot used for the bed.

we tested Geko lab. for tin minerals, and the preconcentration excellent for - 2 mm. So we recommended on the final flowsheet. Hwever the plant was given to a chinese company that change for a kind of Panamerican jig design from the 60'. So the recovery of the total project was less than the design.

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

I have experience with SAG in 90 t/h in gold mine. We could increase throughput to 120t/ h. 

It was work properly. The ore crushed by jaw crusher then fed to SAG. Cyclone overflow directly fed to cyanide tanks.

The crushed ore, after classifying in closed circuit with ball mill, can be routed to flotation circuit where Zinc can be floated at high pH using caustic soda as pH modifier and Xanthate as Collector. Non float is your Tin Ore. The crushing - grinding circuit must be very precise, otherwise whole mass balance will be disturbed.

BUT while a fine suggestion in terms of separation technology, the issue is the difference in grind requirements and, to some extent, the behaviour of the two minerals during grinding.
The cassiterite is coarser and more brittle while the sphalerite is apparently much finer. Chasing the sphalerite initially may well 'bug dust' the cassiterite, making it much harder to recover in any following separation stage.

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

Have a look at our Mina Pirquitas flowsheet, in Northern Argentina, is the most southerly mine in the Bolivia Ag/Zn/Sn mineralized belt. I suspect your project is in the same mineral belt.

Flowsheet is stage crush, Gekko IPJ as a preconcentration step. Ball mill, Ag flotation, Zn flotation. Grind size P80 140-150um.

As full tin recovery circuit was built, dual cycloning to produce sands, fines and discard slimes. Sand and fines conc to dedicated spirals, spiral concs to tables, then a sulphide float for pyrite removal from gravity conc, The discard slimes went to low pH pyrite float, then tin float, followed by multistage cleaning.

Well the tin circuit never operated at an economic recovery.

Due to exactly your problem cassiterite mineralogy, brittle and sulphide contamination of concentrates, thus poor marketability.

Paul Morrow
8 years ago
Paul Morrow 8 years ago

Regarding the energy consumption and Maintenance, I think SAG is not proper option.
Just consider the Top size which is necessary for SAG, you have to design wider Conveyors to be able to handle big particles. Your Chutes and Feeders must also be designed to handle big rocks.
Lubrication System of SAG is another story required special Maintenance.
As your operation is not big, you can not manage large stockpiles, so you will have variability of ore size depending to mining operation. If you feed your SAG by fine particles (less big rocks) you will loss your production because of high circulating load.
If the Silica bearing minerals are higher than normal, you will have more rejected particles required pebble crusher even small ones!
However, if your operation will be developed in future, don't hesitate that SAG is best option.

El Valle Gold Mine in Spain has a Sag MILL since 1998. Commissioning hard due to pebble ports. Performance right. Key is hardness of material.

This mine belongs now to ORVANA. 2000 tpd + the Don Mario Mine in Bolivia also.

Armanis Gold+poly-metallic Mine in Armenia is using Chinese SAG mill 5518 and Ball mill 3245 for grinding, followed by flotation of Lead and Zinc.

Anyway, yes it is possible you are requested to ascertain the facts of comparison of cost benefits [ capex -opex] from Hutti gold mines, Karnataka where both conventional crushing - ball mill circuits and SAG Mill -Ball mill circuit is used for grinding and preparing feed for CIP in their gold plants at the rate of 2000 tpd each.

You know, some time ago you asked about another style of jig for your duty. You might want to take a look at a Kelsey centrifugal jig. My personal experience with this jig was not great, but that was in the early days of its development. Since then a number of metallurgists have advised me that they are working very well.

Helena Russell
8 years ago
Helena Russell 8 years ago

Yes we tested the Kelsey Jig for tin tailings, the pilot test shows that this equipment is the best, so according the capacity of the tin tailings we include or not this equipment. The result shows up grading up to 7 - 12 times, with 70-85% recovery (tin ). Compared with Falcon/Knelson only upgrading 1.2 - 1.5 times for the same recovery.Knelson new model was slightly better than falcon.

Victor Bergman
8 years ago
Victor Bergman 8 years ago

You will be interested in a technical paper I will present at AUSIMM Mill Operators Conference in September in Australia. Gekko have initially been working with HPGR technology for this purpose but have found for the smaller applications that they are both capital intensive and have some inherent operational issues. We found moisture was a considerable concern, especially where you are trying to crush to very fine sizes. Secondly caking of the product which requires special treatment if you wish to screen afterwards and lastly feed preparation in as much as the feed configuration to ensure even feeding to the units.

Gekko have been working with Vertical Shaft Impactors for the last 10 years for all the reasons discussed around fines production and what I would call "un-controlled size reduction". We see the VSI as the perfect tool to liberate minerals sequentially and allow them to be removed from the circuit at their natural liberation size. Energy consumptions as low as 2-4 kWh/t from 40mm down to 600-800um. The CGT / Ballarat Plant VSI runs at 3.9 kWh/t as against the modelled power consumption of a SAG for the same project at 12kWh/t. We see almost NO over-breakage in these plants.

The other conversation is around the IPJ. There has been some great work done as discussed ie. the Mina Pirquitas. I am not sure when the work was done for your project Jorge but just for the record if the IPJ should give the best result for fine jigging down to approx. 75um when things start to fall away fairly significantly. This is if the sp.gr. differential is around 0.8. In other words if the gangue is around 2.7 sp.gr. then we would like to see discreet mineral particles not lower than about 3.5sp.gr. for good separations. Certainly work in the range below 2mm is in the IPJ sweet spot and we should get excellent separations for the lowest operating cost.

The Kelsey jig will not happily work in this range -2mm but likes a finer feed typically below ~0.5mm. I definitely agree it is the best option of the finer fractions and will give the best upgrades but has limited capacity per unit.

Dense media struggles below about 500um as viscosity becomes a problem. Also you would need to run FeSi as the media due to higher density and it could get complex and expensive at these sizes due to media loss etc.

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

I know the Plant Manager at Farallon Mining and it had an AG mill designed to mill 1,500 t/h of ore.

After careful review of the manufacturer specifications Farallon added 1% of steel balls and adjusted the steel balls addition rate and slurry percent solids to reach 2,000 t/h

This (now) SAG mill was still able to hold more steel balls to reach up to 10% steel balls but we didn't go any further as the mine wouldn't be able to mine ore at higher rates.

This SAG mill is working in combination with a tower mill as a secondary grinding stage. Yes, there is no ball mills in that grinding circuit. The p80 to the SAG mill was at about 6" while its product (flotation feed) was at 40 microns.

So the answer to your question is yes, a SAG mill can be installed in an operation milling 2,000 tpd and it can be operated efficiently if the operators have the knowledge to operate it well.

If the ore you are dealing with is a highly competent ore then please don't use SAG mill. Perform competency tests and decide from there.

The biggest problem with a SAG mill is that they are great for a specific ore competency, but that there are very few ores with constant competency. I am yet to find a happy SAG Mill operator who has run the mill for +5 years. I admit to being a 3-stage crush and Ball Mill fan, as this circuit satisfies most ore types (if not too soft).

The 3 stage crush ball mill provides more stability for dealing with different ore types. This circuit can be improved by replacing the ball mill by tower mills, which are more energy efficient than ball mills.
The new advances in technology and manufacturing had led to the construction of more robust crushers and larger tower mills.
the only downside of the 3-stage crush- ball mill (tower mill) is the addition of more screens and conveyor belts which will increase CAPEX and may also increase OPEX.
You noticed that there are more variables in this decision and trade off studies are a must.

Develop a geometallurgical characterization of your ore body to understand ore hardness and competency. Characterize very well the first years of the mine (years 1 to 3 o years 1 to 5) and then characterize the rest of the mine. The first years of the mine will be responsible of most of your NPV.
Model different circuits and then develop trade off studies.
include the CAPEX and OPEX parameter into the evaluation. Don’t design for average hardness-competency but use the percentile 75th or 80th so you design a circuit that will deal with your ore most of the time.
The competency of the ore will be one of your biggest turning points. If the ore is competent then you should avoid the use of SAG mills fed with large rock sizes. In this case the 3-stage crush +ball-tower mill or the 2-stage crush + HPGR + ball-tower mill is your best option
Plants that have competent ore and conventional SAG circuit have reached 50% of design capacity and all of them have needed the addition of secondary crushers stages to reach design capacity.
Some people is now designing SAG circuits for competent ores where they are adding 2 stage of crushing to feed the SAG mill with 45 mm. For me this circuit defeats the purpose of SAG circuits which is to simplify the flowsheet through elimination of conveyor belts and screens.
You have some serious challenge in from of you.

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

A conventional communition circuit with 2 or 3 stages of crushing, rod milling, and ball milling is very forgiving in face of ore hardness variation. One thing, cassiterite has a high specific gravity and will tend to report preferentially to the oversize product of cyclone or screw classifiers. Thus, overgrinding of the cassiterite will be a result of its brittleness as much as of its high specific gravity.

If there is no or limited information about the range of ore hardness, going with a conventional communition circuit would be preferable.

In any event, it does appear that the gravity separation circuit for recovery of the cassiterite will be part of the communition circuit due to the 100 microns P80 needed for the zinc flotation circuit.

If budget and ore supply allows, a pilot plant campaign comparing conventional vs. SAG/BM circuit, with the tin gravitiy in there, and the zinc flotation from the product of the communition circuit.

Please note that - the zinc metallurgy may be better with the SAG as the abraded steel from the grinding media often interferes with selectivity of sulphide minerals. Also note that how tin circuit is implemented could impact the performance of the zinc flotation.

The interactions between the various parts are complex and more easily visible by actual operation (pilot plant) than by numerical simulation.

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

I like your comments because they are aiming put on the table those things that will happen after grinding.

Most of studies will consider the grinding only without paying too much attention to invisible interactions that will happen because of ions load in the flotation area. I agree with you: Zinc metallurgy tends to be difficult due to ions and minerals interaction , presence of ultra fine material, etc.  Ultrafines may be produced due an interaction between the grinding machines, the SG of the minerals present, the knowledge metallurgist-operators have about they circuit (hence their actions on the circuit) and the classification system.

Maya Rothman
8 years ago
Maya Rothman 8 years ago

This is an interesting discussion. 

I agree that SAG have thier place, but in my opinion they are a better fit for simple ore types. 

We all recognise that there will be throughput-and possible product sizing variation, which for complex ore processing circuits adds challenges.

The SAG online time will usually be lower than a conventional circuit, so all downstream equipment has to be sized larger (based upon tonnes per operating hour). This has direct Capex increase. My belief is that this gets 'lost' in the SAG vs crushing/BM trade-off studies.

The very traditional cassiterite hard rock circuits would have stage comminution with recovery processes, repeated in series. Grind-recover grind-recover, etc. Cleary crush and rod mill ball mill-recover, ball mill recover. Then the discusion of screens vs hydrosizers vs hydrocyclones for size splits.

For complex silver-zinc-tin ores, my recommendation is no for SAG.

Victor Bergman
8 years ago
Victor Bergman 8 years ago

Great contributions from all here.

We normally step back after grinding as it is the final P80 that we are focused on.

Creation of too many fines is something that we do work on as "overgrinding" is wasting energy and media for no reason if the lab has determined the size for best recovery to cost.

We all will agree on smaller media = finer grind.

It is for that reason that we encourage porting of smaller media from the SAG and reloading these round rejects to the BM. There is also ways to control the levels of size groups within the BM as well, we are now doing this with blending ball sizes with matched grades of media. This is very useful work as to coarsen the P80 = higher throughput for the circuit.

If there are too many fines or the P80 average reports below targeted size then move up a bit in size of media by way of incremental blending. This is a very controllable environment that does respond to such changes in media.

My thanks to the contributors that are expanding our understanding past the final grind.

I am not a great fan of rod mills after my first hand observation of a serious "tangle".

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

Helena, 

to more clearly/easily assess if SAG milling would be counter-indicated for this tin-zinc application, I would suggest to perform some size-by-size assaying of the material as feed and as product from communition tests - notably Drop Weight/SMC, the ore abrasion test and the Bond Rod Mill WI test.

If the size-by-size distributions of tin and zinc are very close to each other for the various samples representing as-received, "SAG ground by impact", "SAG ground by attrition", and "Rod Mill ground", then it would be unlikely that ultrafine tin not recoverable by gravity separation would be produced from SAG milling. Certainty would only be obtained from pilot testing.

On thing, it might be possible to adapt the GRG test methodology to this material to better characterize the recoverable tin by gravity and at what size to do it more effectively. I was in dialogue with Andre Laplante during the research program for adapting the GRG to PGE metals and from my memory of the discussions we had (with the data on hand) I believe it would be possible to apply the methodology to this case.

Helena Russell
8 years ago
Helena Russell 8 years ago

Thanks to veryone. I agreed that SAG is not the best option in our case. So still evaluate some alternatives. Also Gekko, on this application is not optimum.

Thanks again for your help.

Please join and login to participate and leave a comment.