The main problem of designing the composition of the coal feedstock base for carbonization remains to be the combination of the properties of the components of coal blends. The generalized structural characteristic of coal organic matter (COM), which determines he reactivity in various technological processes, including coking (carbonization) is often considered to be data on the elemental composition of coals [1]. Kiselev and his coworkers [2, 3] have shown that, apart from the elemental composition, for the assessment of the coking power of coals and the designing of the composition of coal blends, it is necessary to take into account data of proximate and petrographic analyses as well. The method of factorial analysis has been used to establish the relationship between the linear combinations of a number of properties of coal blends (so-called main components) and the strength of coke [4]. The main component, which describes the degree of metamorphism of coals, appeared to be most informative.The purpose of this paper is the estimation of the predicting capacity of various criteria of coking power based on sets of properties of coals and coal blends proposed in the literature. The calculations made were based on the petrographic model worked out in Institute for Solid Fossil Fuels (IGI). According to this model, data on the maceral composition of coal components, average values of the vitrinite reflectance index R_0 and plastic layer thickness y are used to find the leaning index i_(lean) and the coking capacity coefficient K_c, which determine the predictable coke residue in the large drum G and other strength indices (M_(25), M_(10)). The algorithm of such calculations has been described by Yeremin and Gagarin [6]. It should be noted that variations of petrographic indices may result in similar values of the elemental composition and other indices. This is responsible for certain ambiguity in predicting the coking properties of both individual coals and especially of their mixes. In other words, the effect of typical composition and property indices on the formation of the reactivity of coals in general and of their coking power in particular is, to a large extent, probabilistic and for the analysis of the role of the composition of coal blends it is expedient to use one of the methods of statistical modelling (for example, the Monte-Carlo method [7]).
展开▼
