Identification of tar-rich coal based on the true density and natural gamma ray response of coals
[Objective]Tar yield,derived from the Gray-King assay,serves as the sole metric for evaluating tar-rich coal.However,insufficient exploration data hinder its effective application in the fine-scale evaluation of extensive tar-rich coal.[Methods]Based on binary classification,this study proposed the optimal thresholds for tar-rich coal identifica-tion using true density(ρ)and natural gamma-ray(GR)log.Moreover,this study elucidated the intrinsic mechanisms be-hind the significant petrophysical parameter responses of tar-rich coal.[Results and Conclusions]Key findings are as follows:(1)Coals with various metamorphic grades differ in identification threshold.For instance,long-flame tar-rich coal,the optimal identification thresholds consist of raw coal's true density of<1.41 g/cm3 and GR responses<80 API units,corresponding to identification accuracy of up to 81.82%.(2)The geophysical log parameter responses of the aforementioned tar-rich coal are affected by inorganic and organic components.The ash content,correlating negatively with the tar yield,acts as the dominant factor influencing the true density of coals,and the clay mineral content indic-ated by the ash composition(Al2O3+SiO2)significantly influences the GR responses of coals.Additionally,macerals,de-termining the tar yield,produce a certain impact on raw coal's true density.(3)On a molecular scale,a lower true dens-ity under a dry ash-free basis(ρdaf)corresponds to richer aliphatic structures in coals,creating more favorable pyrolysis conditions for tar production.Conversely,a higher true density corresponds to higher aromatic structure content,which is more unfavorable for tar production through pyrolysis.Overall,the coupling effects of multiple factors contribute to a strongly negative correlation between the true density and tar yield of coals.These findings can serve as a theoretical guide for evaluating tar-rich coal based on geophysical log parameters.
国家科技期刊平台
NSTL
万方数据
