Coal Preparation » Fine Coal
Thisproject conducted an evaluation of the improvement in the sizeseparation step between 0.1 and 0.4mm, using a 1m diameter classifyingcyclone. This has potential advantages over traditional designs,in reduced capital and maintenance costs, simpler plant layout,and elimination of inefficiencies due to poor distribution.The outcomes from this project will allow coal preparation engineersto evaluate the potential of incorporating large diameter classifyingcyclones into plant circuits, and assess the resulting processingand cost improvements.
The evaluation was conducted at Stratford mine, using the entire flow of minus 1.4 mm fine coal in the plant as the feed. The usual 20% solids content of the feed was reduced to 10% for half of the experiments by halving the feed rate to the plant. Three campaigns were undertaken, with various vortex finder and spigot diameter combinations. The existing fine coal classification circuit of a small cyclone bank and rapped sieve bends was also investigated. Additional flow rate/pressure data points were obtained at Burton and Goonyella coal preparation plants.
Conditions during the trial showed some variations, with pressure drop ranging from 75 to 100 kPa and flow rate 180 to 230 litres/sec. There were some minor variations in the feed, but typical values were about 45% less than 0.100 mm, over 40% floated at 1.3 RD and ash values of about 25%.
A 1 m diameter Multotec cyclone form Ludowici Mineral Processing Equipment was used in this study. It had the following features;
- Involute feed entry
- 0.3m equivalent inlet diameter
- 20o cone angle
- Was inclined at an operating angle of 15o
- Vortex finder diameters of 300, 350 and 390 mm
- Spigot diameters of 160, 180, 200 220 and 240 mm
The feed, oversize and undersize flow rates were about 300, 40 and 260 litres/sec respectively. Unbiased sampling of these flow streams was achieved, and samples were analysed for solids content, ash value and size distribution using both screens and laser sizing techniques. A number of samples were also subjected to particle size/density float sink analysis.
The inevitable errors obtained during the experimental analysis required that the raw data be subjected to mass balancing. The high quality of the primary samples meant that there was little adjustment (less than 1%) required for most of the measured values.
The 1 m cyclone produced acceptable separations between 0.2 and 0.4 mm, and was able to produce the same overall separation size as the combined 490 mm cyclones and rapped sieve bend circuit in use at Stratford. The Rf for the latter circuit was slightly lower than that for the large diameter cyclone.
A model describing the product flow streams as a function of changes in the feed stream or operational variables, to an accuracy necessary for the simulation to serve a useful purpose, was developed. The model requires details of the feed with respect to solids content and washability, and the values for the operating and equipment variables (pressure and vortex finder and spigot diameters). It provides flow rate through the cyclone, amount of water in the oversize stream, partition coefficients for the individual particle size/density fractions and the overall size partition curve.
The density effect, defined by the spread of individual particle density fractions and included as a model parameter, increased with increase in the solids content of the feed. Within the accuracy of measurement, the separation size of the heaviest particles remained relatively constant at 0.08 mm. The lightest particles' separation sizes varied from 0.25 to 0.5 mm depending on operating conditions.
The report concludes that a single, large diameter classifying cyclone can be used rather than a conventional nest of smaller cyclones. It also concludes that sieve bends could be replaced in certain situations.