However, the application and ore characteristics need to be evaluated beforehand. Generally, there are two basic scenarios where a mill sizing can be provided:
Duty specification: Usually this comprises a description of the rock and its comminution characteristics, together with the required tonnage, feed and product size. From this, the type and size of mills for the application are ascertained.
Power demand specification: When some process design work is completed, the specific grinding energy demand is known. With appropriate design information, an operating power demand is converted to a mill size and motor selection.
Duty specification: Before comminution work is undertaken, it is worthwhile getting external advice from providers experienced in the field. They can usually offer initial advice on issues such as what ore types to sample, what testwork to undertake and what variability testing is needed.
Experienced technology providers can advise on suitable laboratories to undertake the testwork and provide the necessary interpretation of the results. Sampling and testing is generally required on any material that makes up more than 10% of the feed material. Perform testing on individual ore types and not a blend of material. Testing of blended material obscures the behaviour of individual material types and does not allow modelling of alternative blends.
The type of data needed is classified as either rock breakage information, useful in crushing and AG/SAG milling, or grinding information for rod, ball and pebble milling. Other useful data is obtainable.
One of the difficulties in evaluating AG or SAG milling is the large pieces of ore or large diameter drill core necessary for an assessment of properties at the coarse size typical of AG/SAG feeds. As a result, tests are applied on smaller-sized material.
These include the MacPhearson test (-32mm) and the SPI test (-25mm). Each relies on a database of operational experience compared to the resultant value in order to predict accurate AG/SAG mill specific energy consumption. The SMC test (quarter core from 50mm diameter core) provides a drop weight index (A x B) to compare with the JK database.
If appearance functions for the JK SimMet software are needed, a number of full drop weight tests are required to validate the results.
The effects of coarse rocks of high competence that do not readily break down on impact are very significant on AG/SAG mill performance. Tests using smaller particle sizes can, on occasions, significantly underestimate the comminution energy distribution between milling stages, leading to rate restrictions in the primary mill. In larger projects, it is always wise to undertake pilot plant testing if possible using a range of feed size distributions to gauge their effect.
The SPI and SMC tests are also applied in operating mines to profile changes in SAG grind-ability within the orebody, which allows refinement of production planning in relation to the energy available in the SAG mill.
If crushing the ore to provide feed to a ball mill circuit or the application is a regrind duty, only the Bond grind-ability work indices are generally needed. This is because the rock is already broken to a small size, usually –12mm, so breakage data is not needed.
To get the most suitable mill sizing for a project requires the following design and operational information:
Throughput – this is reduced to an hourly throughput when the mill is actually operating for correct sizing. Specifying “2Mtpa” is no good unless the availability and use is known
Circuit product size – this is normally called up as the 80% passing size, but sometimes the requirement is for a given topsize or 95% passing size
Ore sources – usually derived from the mine plan, providing information on types and possible throughput changes with time.
AG/SAG mill sizing is generally undertaken using either or both of two methodologies: power-based and breakage-rate based.
Power-based modelling predicts the specific grinding energy needed in the AG/SAG mill by comparing the breakage and grinding indices with those in a database acquired by the tester or technology provider. Breakage-rate-based modelling is done with a software package and suitable breakage parameters.
Each has its limitations and the ideal case is to have enough data for both methods, so that validation is not an issue.
For rod, ball and pebble mills, power modelling is usually sufficient.
A detailed analysis and interpretation of the testwork results leads to a circuit configuration (single or multi-stage) with values for the specific comminution energy demand (kWh/t) in each stage and an estimate of circuit power efficiency.
From the impact characteristics, the minimum mill diameter needed to achieve breakage and the preferred mill aspect ratio (length/diameter) is derived. Next, the ball charge and required ball size are estimated.
If the power efficiency is poor, alternatives such as non-standard feed preparation and the benefit to be obtained from installing a recycle crusher circuit are evaluated.
At this point, the interaction between mechanical and process design aspects arises and consideration of the need for and benefit from a variable speed drive system and design details such as grate aperture are also resolved.
Power design specification: If some design work has been completed and the mill’s operating power (kW) is known, only the mill size needs determining.
A range of options is available in terms of length to diameter ratio (the so-called aspect ratio) and ball charge. To allow the technology provider to specify an appropriate size, either a summary of the ore behaviour or the testwork results is needed, as well as throughput and operating conditions such as open or closed circuit.
Flow considerations are taken into account when sizing high capacity mills operating in closed circuit, with inlet velocity, dynamic residence time and grate discharge flow capacity checks needed.
This is to ensure the correct aspect ratio and feed/discharge arrangements are nominated for the duty.
For an AG or SAG mill, Outokumpu, for example, determines the mill size using a program known as SAGPower, which predicts the power draw of a mill based upon information such as mill diameter, mill EGL, operating speed, ball charge, total mill load, ore SG, milling density, liner thickness and head and grate angle.
Motor size is selected based upon the expected upper operating ball charge and corresponding total load. Once the required dimensions and motor size are known, the mill is then sized, configured, costed and engineered.
Summary
The comminution circuit is often the single most costly part of the process to build and to operate.
So, if you want to get the grinding mill or mills correctly sized firstly, sample the material to be ground without mixing differing ore types.
Secondly, spend time and effort on a comprehensive test program; not forgetting to enlist some expert advice on its content. Then get a reputable technology provider to interpret the data and translate it into a feasible circuit and power demand requirements.
At this point, if you have not already done so, you can approach the technology provider to confirm the mill size that you have, or to provide a mill size for you. The provider will detail its view of the requirement based on the quality of data provided. Poor quality data usually leads to very conservative mill sizes.
The effort applied to characterise the ore will pay dividends in reducing uncertainty in the comminution circuit design adopted, and maybe a little extra production when needed.
Oskar Gustavson has worked as an engineer in the mining industry for 15 years and is global technology manager – grinding for Outokumpu Technology.
More info: (08) 9211 2200