Monday, 25 August 2014

Where are SAG mills going?

In our recent conversation, Prof. Alban Lynch was very sceptical about where semi-autogenous grinding (SAG) mills have gone, and thinks they have taken us along the wrong track. He feels that they may be attractive in terms of high productivity, but says that "their energy efficiency is poor, and I feel that these immense machines are going to turn into very large ball mills".  Developments in fine crushing machines may turn out to be viable alternatives to SAG mills.

Interesting, as in a recent article in Mining Magazine (July/August 2014, pp. 27), Kurt O'Bryan of Weir Minerals says that, although SAG mills have been around for over 40 years, the demand to process more ore in less time is a key factor in where comminution trends are heading.  He says that "SAG mills have to be bigger to meet these demands. Some are up to 12.1-12.8 m in diameter in order to handle the higher tonnages, and the size of these mills makes them harder to install and operate efficiently".

He says that Weir is seeing a rise in demand for larger cone crushers that are matched with large high pressure grinding rolls (HPGRs) for customers who want to replace SAG mills in order to increase efficiency. Utilising cone crushers and HPGRs allows ore to be processed from 250mm to 50mm in cone crushers, then down to less than 6mm from HPGRs. O'Bryan also explains that "the interparticle comminution inherent in the HPGR process is uniquely efficient relative to other forms of comminution in crushing or milling".

So what do the operators think? Do you have problems with SAG mills, and how do you view their future? Is finer crushing a viable alternative? And no doubt the manufacturers will have something to say?


  1. May I slightly rephrase the question, Barry? Can people say that either they are happy that they selected sag mills or say they made wrong selection? In either case, if examples and a bit of information on ore type/top size/hardness of the ore/power consumption etc would be informative, if available and can be shared.
    Rao,T.C., India

  2. SAG mills are doing just fine. Not as well as HPGR's in recent years. The industry has expanded quite nicely in recent years bring upliftment to the countries that have installed them, Zambia, Peru, Chile, Laos and Canada to name a few. SAG mills will be here to stay from some time yet. Understanding how they work and how to model them is probably more of an issue. Proper trade off studies that correctly present the case for either SAG or HPGR need to be considered. Please contact me if you have any questions Regards Mike Daniel. See also

  3. I concur with what Prof. Lynch's comments with regard to his concern about the poor energyefficiency of SAG/ROM-ball mills. I recently came across a paper in Minerals Engineering (Minerals Engineering 24 (2011 1053-1061) that I had previously overlooked. This paper (Energy efficient comminution under high velocity impact fragmentation) demonstrated, in principle, that if particles can be impacted at velocities of a few hundred m/s, the energy to achieve a given grind size can be reduced by a factor of two three. There are comminution machines available that can achieve these velocities based on high-intensity air turbulence. Notable examples are the DevourX and Hi-tec vortex grinders. Some performance data on the DevourX was presented by at the recent MEI Comminution 14 Symposium. However, these technologies, to my knowledge, have received very little attention in the scientific literature. It would be great to persuade the suppliers of these high-intensity devices to scientific investigation.
    Adrian Hinde, Adrian Hinde Consulting, South Africa

    1. My reference to the Hi-Tec vortex grinder was incorrect. Try the following link for a small-scale vortex grinder
      Adrian Hinde

    2. Hi Adrian, who presented the DevourX info at the symposium, you left the name out.
      Thanks, Jason

    3. The DevourX data was contained in a keynote lecture by Alan Muir, which can be found on the Conference Proceedings

  4. Crushers are crushing finer, stirred mills are grinding coarser. Maybe the future is cone crushers-HPGR-stirred mills?
    Keith Pearce, UK

  5. At the tonnages of modern operations and with the gross amount of energy that must be transferred, the key issue is always going to be wear. Autogenous processes have a future. Non-autogenous processes do not. I just recently worked with an operation using VSI crushers (shoe and anvil type) that were going through wear parts in less than a shift (!!). Technically the VSI reduced particle size, but obviously in a non-viable way. I know nothing about the Devour-X, but from the video it certainly looks like there is rock-to-impeller contact and rock-to-wall contact at high velocity, and that alone suggests to me that it has limited viability.

    To some degree FAG's have failed because we can't get enough energy intensity into the ore. Thus the almost universal trend from FAG's to SAG's, along with the migration to ever increasing ball sizes, but even these solutions fail to some degree. Virtually every SAG circuit ends up reincarnated as a SABC, with emphasis on the "C". Crusher energy is not incidental. When you see a circuit designed with 2.5% of the energy coming from a low-tonnage crushing circuit, you can make good money betting that that there will be another capital project in the next five years to raise the energy contribution of "C" to 10%. Again, it is all an issue of energy intensity, which is higher in the crusher but ends up giving you high capital and maintenance intensity per tonne treated.

    The question then is how you get energy intensity into rock without any significant contact between the rock and the machine. A SAG uses low intensity energy storage (lifting) followed by more intense discharge (falling and impact upon the rock "toe"). In that regard it minimizes machine-on-rock contact during the event in which the energy is "consumed". The beauty of solutions such as HPRC is that it has the same effect without being constrained by the limits of gravitational energy storage. Energy is stored in the form of strain on the rocks, which is gradually imparted by the rolls. It discharges in fracture propagation, so that the face of the rolls crusher is taking only a minimal amount of damage.

    We can count on the following - that the emerging processes will be to a large degree autogenous; and that they will impart more energy intensity than we can currently achieve by gravity, since that is clearly a limiting factor in current design. I do think that the day of the SAG is done. The "size" (volume) goes up as the square of the diameter, but the energy intensity (gravity*height) goes up only linearly as a function of diameter. Thus, to double the impact from where we are now you need an 80 foot mill. Compare this to a smaller rolls crusher that uses a different mechanical principle, or an autogenous VSI that is limited only by how quickly you are willing to turn the shaft. Such units allow you to de-couple energy intensity from size. It seems evident to me how this ultimately will end.
    Gregg Hill, Manager Laboratory Services at XPS Consulting & Testwork Services, Canada

  6. A.Bandyopadhyay ( Bandyo)26 August 2014 at 12:39

    It is well an established fact that 3 stage crushing with two stage milling consumes much less power than that by other route. This arrangement got disturbed when treating low grade ores that required significantly high tonnages to achieve economy of operation. There is a limitation in capacity for three stage crushing, around 500-600t/h and thus handling of 800-900 t/h ore throughput was not possible through it. Also the ball mill diameters has reached limiting proportions. Hence AG mills came into picture in a large way and were very quickly modified to SAG mills and depending on the hardness of ore the crushing too was included in the circuit, though of course ABC circuits are not that common. However it must be told that AG mills are no recent development and there are plants in India which has been using AG mills in Gold ore milling for the last 70 years , if not more. I guess , keeping in mind the overall capital cost of new installations SAG mills will remain attractive even though power cost is high. HPGR is of course a viable alternative so far as energy is concerned but in terms of wear of it would be higher because of the very nature of the mechanism. I am not aware if any HPGR installation has been put up to handle 800-900t/h capacity in hard rock situations.

    1. Boddington can do well over 3000tph through each of it's 4 HPGRs on a very hard ore.

  7. SAG mills are quite reliable solution in cases when the mill has to "wash" clay out or "melt" snow and ice coming with ROM ore. As far these 2 factors are not subjcts of human influence - SAG mills have clear and predictable future.
    Vadim Bondarenko, Head of Moscow Representation at Mertec AG, Russian Federation

  8. What are the advantages of SAG mills vs high pressure rollers? I can only think of throughput as the only good point of using SAG, but how do wear and energy use compare between the two?
    Kate Siew, Australia

  9. Barry - thank you for discussing your recent conversation with Alban Lynch. Responding to such a renowned opinion is risky because each one of us knows part of the answer, but no one knows enough to say definitively what the future of SAG mills will be. Discussions posted to date offer many good ideas, but also misinformation. A SAG grinding text book is required to deal with the complexity, the successes, the facts, and the misinformation, which confound students entering the field. This is further obscured by private technologies which rely on private information and computer programs as the basis for new mill designs. The confusion will remain until black box programs delivering modeled solutions are replaced by experienced engineers who use simple ore hardness measurements to calculate the power required, in new single stage SAG and SAG ball operations. The required tools are available and have been proven, but our industry prefers to trust a computer program more than experience. This has to change.

    It is sad but true what Dr. Lynch has said. SAG mills are under-designed, and this occurs because of simulation techniques that pass one error on to another design, and the mistaken philosophy that a SAG mill must be pushed to its limit at all times. A SAG mill can always be pushed further by adding steel and finer pre-crushing. But we have forgotten that SAG milling originally was introduced to eliminate multiple steps of size reduction and do the grinding in a single mill, to reduce capital and operating costs. Because of private technologies, the correct way to design SAG mills has not been questioned by our industry or made a priority at the universities that train new engineers. There are exceptions to this in certain universities who have encouraged SAG mill design lectures to be presented. Instead we embrace a return to the old way where crushing, including HPGR (which is excellent technology), and high steel ball loads, are used to displace the lower cost, operating efficiencies and simplicity that SAG milling offers when properly designed.

    The goal is to generate grinding processes that are economically feasible, environmentally safe to use, as energy efficient as possible considering economics, and which grind design tonnage. Climate will affect mill designs because water shortages can favor the crushing approach while cold weather and freezing conditions will favor SAG milling and its smaller footprint. Processes that reduce the number of units required will help to reduce capital and operating costs and since ore grades will continue to fall, tonnages for economic operation will increase. The real cost of building crushing plants must be properly considered at the design stage. How many people are required to keep the plant clean, and what is the total connected power for the plant including conveyors and dust collectors? Trade-off studies often do not include real assessments of these items. In many cases plant cleanliness is sacrificed to keep the operating costs under control. This kind of operation is unsustainable and either owners will forbid it or regulations will force the change.

    The development of new comminution technologies to replace SAG milling at the tonnages required to treat low grade ores, will take decades. So SAG mills are here to stay. Let's learn how to use them properly. At Starkey & Associates Inc. we are forward thinkers and each of our engineers have given me great ideas about what this posting should say. I thank them for helping. We will be pleased to answer your questions, properly design your new SAG mills and help to optimize existing ones. That is our mission.

  10. “Crushers are crushing finer, stirred mills are grinding coarser. Maybe the future is cone crushers-HPGR-stirred mills” (Keith Pearce, UK). Perhaps the future is not far away. “There is a rise in demand for larger cone crushers that are matched with large HPGRs to increase energy efficiency when crushing from 250 mm to -6 mm” (Kurt O’Bryan Weir Minerals). It is likely that circuits involving explosive breakage, cone crushers, HPGRs and stirred mills will be used for many operations as we respond to the requirement for better energy efficiency. We should not overlook the value of cone crushers. Their high efficiency performance was essential in the operation of the 50 million tonnes per year crushing plant at Bougainville (1969-1989) where crushing to p_80 was in two independent parallel lines of crushers and screens. The plant operated for 96% of the available time and the run time for both the secondary and the tertiary crushers was 91%.

    Gregg Hill (XPS Canada) emphasised the importance of energy intensity, which is high in stirred mills (300 kw/m3) but very low in tumbling mills (20 kw/m3). “The ‘size’ (volume) goes up as the square of the diameter but the energy intensity (gravity *height) goes up only linearly as a function of the diameter”. Little wonder that tumbling mills, which transformed mining when they replaced stamp mills more than 100 years ago, can be described today as the dinosaurs of modern industry. Gregg has set the target with his comment that “emerging processes will be to a large extent autogenous and will impart more energy intensity” although, as John Starkey inferred with his comment that “SAG mills are here to stay”, the changes may take some time to be well established. No surprise about that because major changes in mineral processing have always taken 20 years or longer to be widely accepted and used.

    John has also referred to the need for experienced engineers who use simple hardness measurements to calculate the power required to replace black box programs delivering modelled solutions. I agree, and I have contributed to comminution modelling for 50 years. Modelling may be essential in many technologies because it gives accurate results but in comminution it is only an aid to the experienced engineer. The role of Universities in this area is very important but it is better that I should not go there.

    I learnt much during my conversation with Barry, who is a gracious and very perceptive interviewer, and from the many fine contributions. Thank you.

    Alban Lynch, Brisbane, Australia

  11. Although my experience operating SAG mills is very limited, I feel that the introduction of very large SAG mills has changed operator mentality away from process optimisation (grades and recoveries) to simply maximising throughput (inevitably at the expense of flotation performance). I predict that the next generation will be strongly influenced by this and will continue this trend.

    With diminishing ore grades, I cannot see the SAG mill taking a step backwards any time soon as operations will be forever driven to increasing throughput in order to cost effectively recovery the ever dwindling metal content of the material. I am also not convinced that an alternative currently exists (or if it does exist is trusted by the industry) to very large SAG mills combined with very large tank cells to treat very high throughputs. Okay, recoveries maybe low at 80-85% however reducing throughput in order to increase recovery may well render these operations uneconomical.

    I do believe that pre-grinding beneficiation will become available (or widely accepted) in the future, however this will simply allow the mine to excavate more material while maintaining the same mill throughput.

    My verdict; SAG mills are going nowhere.
    Michael Myllynen, Zone Metallurgist South America at Magotteaux, Chile

  12. Of course in future the comminution will be spend low energy. But how much? Now in industry is the lowest energy in rolling mills. Theoretically, the lowest energy in the microwave mills. If you grind the gold- quartz ore for liberation the gold particles in the microwave hypothetical mill, the specific energy consumption will be 0.08kw/t!
    Size of valuable minerals in the ore is reduced every 10 years. If in 80-s ore more milled to 80% -74micron, in 90-s -44micron, in 0-s -20micron, now grinding limit is reduced to 2micron. Even finer grind impossible. The particle size of 1 micron have 100% plastic deformation. At the same time, geologists found in ores of valuable minerals in size of 10 ... 1nm or even less. How to liberation these minerals? I think it is the task of next generation of comminution. For this necessary known new nature Law for nanoparticle.
    Hvan Alexander, Uzbekistan

  13. SAG designs now are at an all time high in size, I am aware of 14m (46') about to be delivered now.
    The ball makers are also keeping up with new designs that will allow larger media with low breakage in high impact environments.
    There does seem to be a trend of SAG mills being changed into high aspect ball mills, we have encountered a few in recent times. This has brought about a new ball design to reduce spawling as the higher ball charges are causing increased ball on ball incident.
    So perhaps we should not evaluate SAG mills in the traditional configuration of perhaps a +/-12% ball charge and a <20% total charge in a high aspect shell design. Yanacocha for example, has a 19% ball charge and a 21% total charge - Newmont has explored every other possibility with the final result of staying with this loading as it is the only way to meet target.
    The point being that you have a mill that has great versatility, even to deploying it to a single stage milling process. The technique may change, but the machines are probably here to stay...

  14. I think now there is no alternative to the SAG. The main drawback of SAG - high power consumption. To reduce power consumption, I think:
    1. The right determine of SAG feed size. This depends on the mill diameter and the ball size. Feed size necessary to take 80% of the maximum piece of ore that will crushed by the balls. For the big SAG more than 8 meters optimal feed size will be - 50mm. Ball size - 120 ... 125mm. Do not a pebble crushing , the need secondary crushing.

    2. Reduce the power consumption for the drum rotation of the mill. The power consumption in the mill is distributed between the drum rotating and moving the balls. The first harmful, second useful.
    For example: For the drum rotation we have a power consumption 800 kW / h. If we load balls, the power consumption increased to 1600 kW/h. Here 800kW/h is for moving the balls. Grinding the ore will be only 800 kW. May be the drum manufacturing from polyurethane with the lining. The power consumption for rotation of drum will be reduce to 400 kW and all power consumption reduce to 1200 kW or 25%. For roller mills polyurethane impossible.

    3. It is necessary to increase the of ball charge up to 25%.

    4. It is necessary to increase the power motors of all the SAG application now.

    5 It is possible to increase the diameter of SAG but now there is no need. SAG of 12 meters and 20% ball charge can be achieved capacity up to 10000 t/h.

    Hvan Alexander, Uzbekistan

  15. I am really happy this Blog and the others are being actively participated. It is very encouraging to the future generation of Mineral Engineers, particularly students, to see the challenges which lie in front of them for innovation and top class R&D.
    Congratulations and thanks to you and the participants; if others do not misunderstand, I wish to add that I hope to see more R&D experts participate and guide these divergent views -- this is in fact my humble request.
    Rao,T.C., India


If you have difficulty posting a comment, please email the comment, and any photos that you might like to add, to and I will submit on your behalf