The spherical agglomeration process provides an attractive method for the cleaning and recovery of fine coals in the form of compact oil-bonded pellets. In treating a washery effluent containing 50% minus 20 micron material by this technique, the slurry was first mixed under vigorous agitation with oil to deash and dewater the fine coal. The flocculated coal pulp was then formed into larger agglomerates on a modified balling disc. The operation of a semi-pilot scale process is described and an analysis of some of the operating and system variables is given. A semi-theoretical expression to predict the final agglomerate size under various conditions is derived and compared with the experimental data.
Fine Coal Treatment Processes
Most coal cleaning methods depend upon the density difference between coal and its impurities to effect separation. These gravity concentration methods, however, are not practical for particles finer than about 100 mesh and cleaning methods dependent upon the differences in the surface chemistry of coal and foreign matter are used for the finer sizes.
Whereas surface treatment chemicals are used in very small quantities in the flotation of coal, these extremely fine sizes may be removed from suspension only if quite large quantities of oil (5 to 50% of the solids feed) are agitated with the slurry. Two such bulk oil processes have previously been described. In the Trent process powdered coal, water and about 30% oil are beat together to form an “amalgam” of cleaned coal. In the Convertol process 3 to 10% oil is mixed with the slurry under vigorous agitation and the product is discharged directly to a high speed screen (60 to 80 mesh) centrifuge for dewatering. These bulk oil processes produce fine, flocculated concentrates and an important aspect is the question of the dewatering and disposal of this material. For example, one disadvantage of the Convertol process is loss of coal recovery due to a gradual increase in the size of the centrifuge screen perforations caused by wear after relatively small throughputs.
A number of batch tests were performed in a high speed blender (capacity 1000 ml.) to provide data for the design of a continuous process. Variables studied included the type of oil and concentration and level of agitation and contact time in the mixer. The coal slurry to be treated was agitated for approximately one minute and the required amount of oil was then added. Agitation was continued for a further 10 minutes and the mixture was then poured onto a 100 mesh screen to allow the water containing the ash component to drain through while retaining the agglomerated coal. The flocculated product was washed with 500 ml. of water and dried.
All experiments were performed with a 10% W/V coal slurry since this concentration was considered typical of a wash plant effluent. The coal slurry was stored, under agitation, in a 45 gal. drum and was pumped together with 28% Varsol based on the weight of dry coal to a mixing vessel. The mixing vessel was fabricated from stainless steel and was 9 in. diameter by 8½ in. high with a capacity of 8.8 1. to the overflow spout. Coal slurry was normally pumped at 500 ml. per minute leading to an average residence time in the mixer of 17 to 18 minutes. Since oil/coal contacting was complete in about 5 minutes, it is felt that the slurry feed rate could be, at least, doubled with no effect on product grade and recovery.
During these continuous runs, combustible recovery was in the range of 83% to 91%, ash content was 5-6% for the pellets and 60%-70% for the underflow from the float-sink tank. Pellets ¼ in. diameter held approximately 12% surface moisture leaving the disc and air dried overnight to less than ½% moisture.