Relevance of Maceral Concentrates to Whole Coal Coking Predictions

Following on from earlier projects, C21059, C24055 and C25052, this project set out to use the reflux classifier to separate macerals and CGA to identify the structural components of the separated coal particles; and study the impact of these macerals concentrates and their blends on fluidity during coking.

The objectives of the project were to:

• Clarify how to concentrate maceral components from coke oven feed particles combining the reflux classifier and Coal Grain Analysis;
• Examine what maceral concentrates represent in whole coal and the extent of the concentrates and the blends modify industrial fluidity of whole coal; and
• Establish the relevance of coal maceral concentrates on whole coal coking prediction.

Three coals with different ranks were selected for CGA prior to maceral separation. The results indicated that maceral compositions and populations (content as percentage) of these coals varied with coal particles, and the variation was different from coal to coal. Overall, for a single coal, the small particles had higher vitrinite contents than the large particles. Fine particles had a much higher proportion of grains containing highly concentrated-vitrinite (65-95%) than coarse particles, while coarse particles had a higher proportion of highly concentrated-inertinite (65-95%). Two size cuts of particles from each coal were selected for maceral separation. The reflux classifier produced products of lower density through to higher density, in effect concentrating macerals from high vitrinite to high inertinite. About 10% concentrates (by mass) were highly concentrated vitrinite and inertinite rich particles. Because coal maceral compositions are related to mineral content, particles that had similar ash contents with the raw coal were selected as heterogeneous particles along with the highly concentrated vitrinite and inertinite rich particles for CGA. The CGA results showed that the RC methodology could substantially shift the maceral distribution from an evenly mixed stream of high purity and dominant grain vitrinite (low flow rates) to streams containing an increasing amount of inertinite as dominant grain and composite grain (higher flow rates). Overall, the CGA results for the RC products indicated that the reflux classifier was able to separate the coal into a highly concentrated vitrinite product and inertinite rich stream regardless of feed vitrinite content. In practise, this mainly impacted the yield of products, rather than quality.

The Gieseler fluidity tests were undertaken to determine the coking impact of shifting the maceral distribution. Blends of “high purity” vitrinite and “concentrated inertinite” were produced to match the total maceral content in the heterogeneous product (from intermediate RC flow rate). The results suggested that the blends tend to decrease softening point and alter plastic range (2-7°C) and show higher fluidity than the heterogenous concentrates. However, the increase of fluidity for the blends seems not significant considering the accuracy of the fluidity test itself (repeatability = 0.1 log MF and Reproducibility = 0.3 log MF).

The fluidity of the maceral matched blends of the vitrinite and inertinite concentrates was also compared against the 2-component polymer blending prediction. The comparison indicated that this blending rule did not produce a predictive value of fluidity close to the measured value, nor provide an adequate trend in its differences. In past studies, this difference between predicted and measured values has been claimed to be a result of interaction. However, it is suggested in the outcomes of this project that an alternative reason may be that the polymer blending rule does not adequately account of solids such as inerts.