Metallurgical Accounting

Production and Metallurgical Accounting
In order to be able to assess the production and metallurgical performance of a processing plant accurately a system of metallurgical accounting is required. This will normally consist of:

Sampling of feed, concentrate and tailings streams
Mass flow measurement of the same streams
Laboratory analysis of the samples
Imputation of the assay and tonnage results into an accounting database (or spreadsheet) to determine grades, recoveries and production.

Based on the information generated from the above, the metallurgist should be able to make necessary adjustments if required and possible to achieve desired plant performance. It follows that the accounting information should be as accurate as possible for this to be done.
Sampling
Samples collected should be as representative as possible of the stream that is being sampled. Samples are usually collected from conveyors (dry) and pipe lines (both wet and dry).

Whenever possible automatic sampling should be used, to eliminate the variability inherent with manual sampling. There are three types of automatic sampler in common use, cutter (cuts straight across stream), vezin (rotates through stream) and deflection (diverts stream).
It is typical to collect 12 hour composite samples to cover the duration of the shifts, consisting of regular multiple cuts.
During a 12 hour shift a cutting frequency of every 30 minutes would be typical, resulting in a total of 24 cuts.
The volume and weight collected during each cut needs to be such that a reasonable sized shift composite is collected for removal and transport from the plant to the laboratory. In order to generate a composite of 2.5 to 7.5 kg this would mean 10 to 30gm per cut. This should then result in suitable sample sizes to be collected in buckets or calico bags, depending on the SG of the sample material.
With cutter and vezin type samplers, the width of the cutter opening and the speed that the cutter traverses the stream will control the amount of sample collected per cut. Both should be adjustable. With deflection type samplers, the speed and time that the deflector is in the open position controls the amount.
Contamination is often an issue, and for this reason buckets should be covered with lids, and calico bags tied to sampling lines. The location of collection points should where possible be selected for lack of overhead spillage.
The advantage of calico bags over buckets for collection of slurry samples is that provided slimes are not excessive, or require assaying, they will usually automatically dewater the collected sample, while buckets must not be allowed to overflow, or splash during transport to avoid solids losses, or be manually dewatered.
Sampling is normally carried out on vertical discharge streams, where the cutter moves across the entire stream. In the case of large slurry flows, this is often a secondary stream to reduce the volumetric flow being sampled.
Best practise is to do this close to the discharge of the slurry pump where flow is turbulent, and on the first vertical section to avoid stratification. A rubber or urethane lined T piece, connected to a rubber lined isolation valve followed by a 25 to 50mm line to discharge in closed circuit with the feed sump, with an automatic sampler at the discharge point is typical. It may be necessary to reduce the volume and pressure by installing a multiway static distributor ahead of the sampler, with only one discharge line being sampled.

Mass Flow measurement
Mass flow is usually measured on dry flows using weightometers (for conveyors), or impact weighers (for vertically falling dry flows), or wet flows using nuclear density gauges integrated with magnetic flow meters (for slurry pipelines).

Weightometers are based on load cells which respond to pressure bearing on rollers supporting a section of a moving conveyor which are mounted on a suspended frame. Double roller weightometers are more accurate than single roller types. The load cells are calibrated by lying chains of known weight on the length of empty belt where the suspended frame is mounted. The resulting weights measured on the moving belt are integrated with a speed sensor on a roller to give mass flow rate.
Impact weighers are based on the pressure exerted by a vertically falling stream of sand sized dry material from a constant height onto an angled plate which is mounted on a load cell. The load cell is calibrated by zeroing for no flow and then collecting and weighing material that has flowed over the plate over a measured time. They are commonly used in mineral sands dry mills.
Nuclear density gauges are fitted to slurry pipelines on vertical sections close to pump discharges (for the same reason as sampling points), and are calibrated with water flowing through the line and then by collecting samples of slurry and physically measuring the slurry density of the samples (using Marcy scales or weighing known volumes of in buckets).
Magnetic flow meters are fitted to similar sized spool pieces of slurry lines on pump discharges (but not necessarily as close as possible or on vertical sections) and are pre- calibrated by the manufacturer. They depend on the motion of a conducting medium through a magnetic field to generate a small electric current.

Laboratory analysis
Sample analysis consists of sample preparation of the samples received from the plant and then assaying.

If a sample is wet then it will need to be dewatered and then dried. Dewatering is typically done using a pressure filter, the resulting damp solids being dried in an oven. If de-sliming is required, this is carried out before dewatering by wet screening at 38um, collecting both the oversize and undersize, the latter collecting in a drum and decanting before drying and weighing.
The dry samples are then reduced in size, typically using riffles or a rotary divider, so that they are suitable for assay. In the case of XRF analysis this is only a few grams, after which the sub sample is ground to a fine powder in a pulveriser, before either being pressed into tablets or heated in a flux to beads ahead of analysis.

Accounting procedures
The use of a database (such as access) to store and calculate results is preferable to a spreadsheet, for security, reliability and input / output reasons.

Tonnage and running time data can be entered automatically by the plant PLC, and assay data from the laboratory computer. This data can then be used to report and calculate feed and concentrate grades, tonnages and rates and recoveries to concentrate.
In many instances where there are only two products, concentrate and tails, and the tonnage of concentrate is small compared to feed and tails tonnage the two product formula can be used to calculate recovery.
This is an assay only based formula namely recovery % = (f x (c-t)) / (c x (f-t)) x 100, where f = feed grade; c = cons grade; t= tails grade.
This can be reconciled with tonnage unit based recoveries where concentrate and feed assays and tonnages are used to calculate recovery, namely recovery % = (c x C) / (f x F) x 100, where C = cons tonnage, F = feed tonnage
The two product assay method typically is more reliable than the tonnage based method over short periods of time eg shift and daily results. Both methods need reconciling with mine production and final shipped product production on a monthly basis. This also involves stockpile (surveyed) and stored product reconciliation.