Improving coke strength prediction using automated coal petrography

A range of Australian and overseas coals with a mean maximum vitrinite reflectance (Ro, max) range of 0.68-1.71% were carbonized in a test coke oven. Coal properties were characterized using XRF, XRD and automated imaging of polished sections of discrete coal grains. The Full Maceral Reflectance (FMR) parameter was calculated from the reflectance data and was correlated with coal as well as coke properties. The Ro, max values and the vitrinite estimates from the semi-automated microscopic technique indicated a good correlation with similar data based on manual point count analysis. The FMR parameter is shown to increase with increasing carbon content, structural ordering of carbon and decrease with increasing volatile matter. The FMR parameter of coal was related to cold coke strength DI15150 and coke strength after reaction (CSR). The FMR parameter was modified by diluting the contribution of high reflectance coal grains as well as incorporating the effect of ash contribution to propose a combined coal index (CCI). The new coal index is shown to improve the accuracy of coke strength prediction. The combined coal index provides a promising objective measurement based alternative for predicting coke strength.