Comminution
UCS, CWI, RWI, Ai, SWI, A, B, Ta
- What do these indices mean?
- How are they determined?
- How much sample do you need?
In most resource projects the majority of the capital expenditure will be spent in the comminution area with the milling circuit itself being the most expensive item. Should the design parameters for comminution plant not be correctly determined then the entire project may fail. It is critically important therefore to have a clear understanding of the number of comminution tests required and to ensure they are carried out correctly on a sufficient number of variability samples to adequately describe the processing performance of all sections of the ore body.
Unconfined Compressive Strength (UCS - Mpa)
The term is self descriptive in that selected rock specimens are subjected to increasing compressive forces using a standard apparatus, until rock fracture occurs. This parameter gives the first indication of competency, and high values will indicate that robust primary crushers must be selected, for example, if a jaw crusher is to be used then a double toggle unit would be chosen.
As an indication of competency those ores exhibiting UCS values in excess of 180 Mpa would be very competent and in addition to the requirement for a robust primary crusher a doubt would be raised as to the SAG mill amenability of that ore.
UCS can be determined on core, or samples drilled from rock specimens.
Bond Crushing Work Index (CWI)
This index describes the competency of the ore at larger particle sizes. It is used to calculate actual crusher power requirements, however if the index is significantly higher than Rod or Ball work indices, this is again a cause for concern if SAG milling is part of the process flowsheet.
In this test twenty representative rock specimens in the size range -76+50mm are broken under the impact of twin pendulums. The input energy is increased until rock breakage occurs and the Bond Crushing Energy Eb is related to a constant for the apparatus and the angle through which the twin pendulums fall by the relationship:
Eb = 82*[1-cos(a)]
The Bond crushing work index is then calculated from the formula developed by F.W. Bond
CWI = [2.59*1.1*Eb/d/sg] kWh/t
Where:
Eb = Bond crushing energy for an individual rock
K = 82, a constant
(a) = the angle through which the pendulums fall
sg = the specific gravity of the individual rock
d = the thickness of the rock specimen
Twenty pieces of representative rock passing a 76mm square opening but retained on a 50mm square opening are required. Values of Bond crushing work index will vary from a 8 kWh/t for laterite hardcap through to 22 kWh/t for banded iron formation to 46 kWh/t for fresh greenstone.
Bond Abrasion Index (Ai)
This index, devised by F.C. Bond in the 1940's, quantifies the abrasivity of an ore. The index can be used to calculate metal wear rates in crushers and ball consumption rates in ball mills.
In this test procedure 4 * 400g sub- samples of ore are stage crushed and sized into the range -19.0+12.7mm. A standard weighted test paddle and enclosure are used and the paddle is abraded by rotation in contact with the ore sample for 15 minutes at 632 rpm. This procedure is repeated four times and on completion the paddle is re-weighed and the loss in weight in grams is the abrasion index.
10 kg of -55+38mm representative ore is required. This is stage crushed to 19mm and screened at +12.5mm and the four * 400g sub-samples in this range are extracted. Bond abrasion indices (Ai's) vary from a low of 0.026 for limestone through 0.18 for quartz and 0.25 for magnetite to 0.69 for quartzite and taconite. Ai's in excess of 1 have been experienced in the gold industry with resulting crusher wear part life of less than three weeks.
Bond Rod Mill Work Index (RWI)
This index is used to calculate the power draw requirements of a rod mill if one is included in the process flowsheet. Additionally it allows more precise calculation of comminution energy requirements. In the calculation of the total energy requirement using the bond procedure the RWI is used for secondary crusher product size say 25mm down to 2100µm and the Bond ball mill work index is used from 2100µm down to final product size.
The RWI allows further observation of the behaviour of the ore at larger particle sizes. An ore with a significantly higher RWI than BWI shows a tendency to be more competent at larger particle sizes and may indicate problems in SAG milling.
In the determination of RWI 12 kg of representative ore is staged crushed to 100% -12.5mm. A standard volume is added to a standard RWI mill and ground dry at 46 rpm. The product is sized on a closing screen selected to be close to the design rod mill product size which will be in the range +208 - 4700µ and is typically 2100µ. New feed is added to replace the screen undersize and the procedure is continued until a 100% circulating load builds up.
The rod mill grindability Grp is the average of the last three cycles product expressed as grams per revolution. The feed and product are also sized and the 80% passing size established in microns. The Bond rod mill work index (RWI) is then given by:
Where:
Grp = Rod mill grindability in g/revolution
P = Product P80 in microns
F = Feed P80 in microns
Pi = Opening size in microns of the sieve size used
10 kg of +12.7mm representative ore is stage crushed to 100% -12.7mm and a standard volume is introduced to the standard rod mill apparatus.
Bond Ball Mill Work Index (BWI)
This parameter controls the calculation of the basic energy requirements of the comminution circuit and as such is the most important parameter. In the determination of BWI 15 kg of representative ore at 100% +3.35mm is stage crushed to 100% -3.35mm. In a similar manner to the RWI a standard volumetric charge is placed in a standard mill and dry ground. A closing screen is selected to target the desired product size and grinding cycles are continued until a 250% circulating load is achieved. The average of the last three net grams per revolution Gbp is the ball mill grindability. The Bond mill work index BWI is calculated from:
Where:
Gbp = Ball mill grindability in g/revolution
P = Product P80 in microns
F = Feed P80 in microns
Pi = Opening size in microns of the sieve size used
15 kg of representative diamond drill core, cut or uncut, is required so that it can be stage crushed to 100% -3.35mm. Typical values of Bond ball mill work index for soft oxidised ore bodies lie in the range 5-10 kWh/t at a standard product size of 80% passing 75µm. As competency rises medium primary ores will have BWI's in the range 10-15 kWh/t and very hard primary ores will be in the range 15-25 kWh/t.
SAG Milling Testwork
A detailed knowledge of a particular ore body together with informed examination of the drill cores will allow a preliminary judgement to be made on the suitability of SAG milling as a comminution option. Ore derived from a very competent uniform ore body will not break in a SAG mill. By contrast a completely oxidised clayey ore body will not be capable of supplying any of the media required for SAG milling. If SAG milling is selected as a possible option ie:
- Unconfined Compressive Strength greater than 180 Mpa
- Bond Crushing Work Index greater than 20 kWh/t
- Bond Rod Mill Work Index is not significantly higher than Bond Ball Mill Work Index and both are not significantly higher than 15 kWh/t;
then further SAG mill testwork can be carried out as follows:
Determination of Julius Kruitschnitt Mineral Research Centre (JKMRC) SAG mill parameters A, B and Ta. AMMTEC has made an agreement with JKMRC and its subsidiary J.K. Tech to carry out comminution testwork using the second generation drop weight tester. This agreement secures exclusive Western Australian rights for this testwork for AMMTEC. The JK drop weight test method provides ore specific parameters for use in the JKSimMet Processing Simulator software. In JKSimMet, these parameters are combined with equipment details and operating conditions to analyse and/or predict SAG/auto mill performance. The same test procedure also provides ore type characterisation for the JKSimMet crusher model.
The drop weight test is used to calculate the energy that is expanded in breaking the particle:
E1 = Mg(h -XM ) (1)
Where:
E1 = Energy used for breakage
M = Drop weight mass
g = Gravitational constant
h = Initial height of drop weight above the anvil
XM = Final weight of the drop weight above the anvil
Sample requirement is 70 kg of representative broken rock in the size range -76 +6mm or 70 kg of whole uncut drill core at a diameter greater than 63mm. The sample is control crushed to generate five size ranges of rock fragments:
- -63.0mm +53.0mm
- -45.0mm +37.5mm
- -31.5mm +26.5mm
- -22.4mm +19.0mm
- -16.0mm +13.2mm
The impact energy is chosen to suit the hardness of the particular ore being tested. The broken ore is collected after impact and sized, and the size distribution normalised with respect to the original particle size. The distribution is described by a single number T10, which is the percentage passing one tenth of the original particle size.
The drop-weight test is carried out on ore particles at five narrow size ranges of particles, each size range at three different energy levels, thus making a total of 15 size energy combinations.
Low energy breakage is characterised by tumbling an ore sample in a small mill running at 70% of critical speed. After 10 minutes the charge is sized and the abrasion parameter, ta, calculated.
| Ore Type | A | B | Ta |
| Soft | 92 | 9 | 4 |
| Average | 56 | 1.95 | 1 |
| Hard | 34 | 0.3 | 0.2 |
Determination of the Autogenous Work Index Using
the MacPherson Procedure
A.R. MacPherson of Ontario, Canada, has developed a test procedure for determination of the Autogenous Milling Work Index. The system uses a 457mm Aerofall air swept mill and is operated for a period of time sufficient to have established balanced mill conditions, and then operated for 1-2 hours during which time samples are taken for size analysis and product weight distribution.
Sample requirement is 227 kg of representative uncrushed ore or diamond drill core. The sample is stage crushed to 100% -32mm prior to testing.
Increasing Competency at Large Particle Size
A typical SAG milling installation will feed primary crushed material to the SAG mill and the 80% passing size may be as high as 150mm with 20% of the rock fragments greater than 150mm. Since both the JKMRC and the MacPherson procedures stage crush the ore sample to -32mm prior to testwork, those ores which show extreme competency in the larger particle sizes may not give a valid result.
In order to be completely sure that the ore is amenable to SAG milling the Advanced Media Competency Test is required. This test is used to assess the suitability of an ore to grinding in an autogenious mill. The ore sample, as defined below, is tumbled in a 1.83m diameter by 0.31m long mill for 500 revolutions at 26 rpm. The product is sized and the number of rock pieces in each size fraction coarser than 19mm is determined.
The size analysis of the mill product can be compared with the corresponding size analysis of ores which are known to be suitable for autogenous or semi autogenous milling. Impact and Crushing work index tests are performed on 20 rocks in each of the following size ranges:
-102+76, -76+51, -51+38, -38+25, -25+19mm
Bond rod mill work index, Bond ball mill work index and Bond abrasion index are carried out on sub samples cut from the main tumbling test sample.
If drill core is being provided the requirement is 200 kg of whole PQ core (85mm diameter). If HQ core (63.5mm diameter) is supplied again 200kg is required, however the crushing work index of the top particle size cannot be determined in this case. If whole rock pieces are available, eg selected from the CV1 on an operating mine, then 11 specimen rocks in each of the size ranges listed above is required.
Ultra Fine Grinding
AMMTEC has available a pilot Metprotech stirred mill for the determination of energy requirements and mill sizes to enter the sub twenty micron area. The Metprotech stirred mill arose out of a research project at the Council for Mineral Technology (Mintek) in South Africa in 1982. The claimed advantages of this mill are as follows:
Low speed stirring action allows the use of small diameter (6mm and less) which gives the optimum angle of nip required for efficient ultra-fine grinding.
The mill operation is relatively insensitive to pulp viscosity, unlike tumbling or centrifugal mills.
The power intensity in tumbling mills is limited to about 30 kWh/t which means very large machines are needed for fine grinding. In the Metprotech stirred mill, higher power intensities are achieved due to the torsional and compressive forces which give the required abrasive action.
The milled products from the Metprotech process give higher gold recoveries than products from other mills ground to the same or smaller size. The reason for this is the higher surface area of the products from the Metprotech process.
The design of the mill allows leaching to take place during grinding. Often, substantial leach recovery is possible under milder conditions leading to an overall improvement in leach recovery. The heat produced during milling improves the dissolution, while the freshly liberated gold is attacked by the leach solution before its surface can coat or passivate.
The ultrafine milling process has potential application in the treatment of both sulphide concentrates and calcines and in hydrometallurgical treatment of base metal concentrates. Fine grinding to produce high surface areas can often mean that leaching is possible under milder conditions thus it may be possible to achieve reasonable leach rates and recoveries without the use of high temperatures and pressures. Coupled with this possibility is the development of carbon and resin-in-pulp technology which obviates the requirement for solids/liquid separation prior to precipitation of metal values from solution.
Two kg of representative sample are required. The raw data is obtained by AMMTEC and Metprotech provide the interpretation and scale-up.
SVEDALA Detritor Fine Grinding Mill
AMMTEC recently took delivery of a SVEDALA Detritor grinding mill. The mill can grind to as fine as 5µm. This ultra fine grinding mill uses silica sand as a grinding media. The use of the cheap media has the potential to significantly reduce grinding costs. In applications where media contamination of the sample can be deleterious to later processing (eg. cyanide leaching), the use of an inert media has significant benefits.
The SVEDALA Detritor is fully instrumented and this allows for energy consumption calculations to be made and assist in the determination of scale-up to full production units.
Pressure Oxidation
The aqueous oxidation of metal sulphides is characterised by complex chemistry with several reactions being possible, either in series or in parallel, depending on the physcio-chemical conditions of the system.
Sulphide sulphur can report in several forms, eg. elemental sulphur, sulphate, basic sulphate, acid or jarosite. The metals report as cations associated with sulphates (or jarosite) or as the oxide. Arsenic (a metalloid) behaves in a similar manner to sulphur forming arsenious acid, arsenic acid, arsenite (111) or arsenate (V).
At ambient conditions the oxidation rate of sulphides (eg. pyrite and arsenopyrite) is low. Rates are usually chemically controlled, although after extended times a layer of elemental sulphur can build up on the surface causing diffusion through this layer to become rate limiting. Vigorous agitation often assists by scouring this layer.
Bacteria can catalyse the oxidation reactions under ambient or 'just above' ambient conditions. A number of mechanisms of catalysis are possible; eg. by direct enzyme catalysis of electron transfer (redox) reactions on metal and sulphur at the mineral surface, and/or indirectly by oxidising ferrous to ferric iron in solution. (resulting in the ferric iron subsequently leaching the mineral). Bacterial oxidation can often become limited by oxygen availability; ie. the chemical reactions are catalysed, but the availability of oxidant becomes the problem, due mainly to the low solubility of oxygen in water under ambient conditions. Thus mass transfer of oxygen is an important consideration in bacterial oxidation.
Oxidation rates of sulphides can also be substantially increased by the application of heat and pressure. Increase in temperature has a marked effect on the rates of reactions with high activation energies according to the Arrhenius relationship:
Rate a ke -E/RT
| Where: | k is a rate constant | |||
| E is activation energy | ||||
| R is the gas constant | ||||
| T is absolute temperature |
The other important effect of pressure oxidation can best be appreciated by considering variation of oxygen solubility with pressure and temperature. To 200 degrees Celsius there is almost a linear increase of oxygen solubility with pressure, however above 200 degrees a very rapid increase is observed. The net effect of increased oxygen concentration at elevated temperatures is to increase its availability at the reaction sites.
For the oxidation of pyrite it has been shown that:
Rate a P(O2)n
Where P(O2) is the oxygen pressure and n lies between 0.5 and 1.0 The value of n would seem to depend on the mineralogy of the pyrite itself.
For arsenopyrite the value of n is unity; ie arsenopyrite oxidation is more responsive to oxygen overpressure.
In general rate of oxidation may be expressed as follows:
Rate of Oxidation = k e-E/RT P(O2)n
Pyrite oxidation is generally considered to take place according to the following reaction:
FeS2 + 7/2O2 + H2O -> FeSO4 + H2SO4 (1)
FeS2 + 2O2 -> FeSO4 + Sº (2)
Above the melting point of sulphur (110ºC) reaction (1) is dominant, whilst above 165ºC reaction (2) is negligible.
Ferrous (11) sulphate is subsequently oxidised to ferric (111) sulphate according to the following reaction:
2FeSO4 + 1/2O2 + H2SO4 -> Fe2(SO4)3 + H2O (3)
Ferric iron can further catalyse the oxidation process and can also be hydrolysed according to:
Fe2(SO4)3 + 3H2O -> Fe2O3 + 3H2SO4 (4)
(haematite)
Fe2(SO4)3 + 2H2O -> 2FeOHSO4 + H2SO4 (5)
(basic sulphate)
Reaction (5) is favoured by high acid concentrations. Reaction (4) is the preferred one in commercial practice, since the residue of haematite is easier to neutralise and cyanide for gold recovery.
Sulphuric acid concentration is therefore controlled to enhance reaction (4); eg it is kept below 70 g/l at 200ºC.
Inert salts, like MgSO4, can further help control the reactions by allowing operation at higher acid concentration whilst still promoting reaction. (4)
Arsenopyrite oxidation takes place according to the following reaction:
2FeAsS + 13/2O2 + 3H2O -> 2H3As)4 + 2FeSO4 (6)
2FeAsSb + 7/202 + H2O -> 2H3As)4 + 2FeSO4 + 2Sº (7)
The reactions are competitive parallel reactions with reaction (6) being favoured by high temperature operation.
Ferrous (11) iron is oxidised to ferric (111) iron as before and the ferric iron reacts with arsenic acid to precipitate ferric arsenate (scorodite) according to:
Fe2(SO4)3 + 2H3AsO4 -> 2FeSO4*2H2O 3H2SO4 (8)
