Jordan, Pierre2008-04-152008-04-152008-04-15http://hdl.handle.net/10539/4761Given the high costs and general shortage of coking coals on the domestic and international markets, and because the nature and qualities of many of the coking coals available on the markets are themselves mixed products, conventional mechanisms and tried and trusted formulae for manufacturing coke products based on single coals of known qualities can no longer apply. There is therefore an urgent need to develop more effective techniques for evaluating and assessing the properties of individual coals rapidly and reliably and in a manner that could provide useful data for use in modelling the effect of new coal components in a coke blend. Towards this end, the current research has sought to find more accurate coal characterisation techniques at laboratory scale than currently exists in industry at present. Seventeen coking or blend coking coals from widely different sources were selected and cokes were produced from them in as close to full scale conventional conditions as possible. Both coals and cokes were analysed using conventional chemical, physical, petrographic and rheological coking methods. The results indicated that, whilst all coals had acceptable chemical, physical and petrographic properties as evaluated on individual parameters thereby indicating their potential values as prime coking coals, in fact the resultant cokes of some of the coals had properties that disproved this assessment. These anomalies were investigated by integrating all characteristics and statistically evaluating them. The result [outcome] indicated that the series of coals under review fall naturally into three distinct categories according to rank, as determined by the reflectance of vitrinite, and that the coking coals in each rank category were further characterised by parameters specific to that level of rank. In this way more accurate predictions of coke quality were obtained than has been the case to date when using single set evaluations or previously devised formulae. On this basis it was concluded that, when selecting coals for coke making, it is essential to first establish the rank of the coal by vitrinite reflectance and then to apply coke evaluating parameters specific to that level of rank. The formulae developed for this purpose held good for all coals tested, however, it remains to be seen whether this applies universally to an even wider source of coals.1367185 bytesapplication/pdfencoalcokecoal-petrographycoke-petrographycoke quality predictionThesis