Technical Market Support » Metallurgical Coal
This project focussed on revealing differences between textures and structures of Northern Hemisphere and Australian cokes made from coals with similar rank (RMax) and amount of vitrinite.
To perform the project, 19 coke samples made from single coal, made from Northern Hemisphere coals, were matched with 29 Australian coke samples previously characterised for project C25048. From seven matched pairs of cokes made from medium/high rank coals, two were matched with the same Australian coke. From six matched pairs of cokes made from low rank coals, four were matched with one Australian coke and two with another.
The actual size of each MosaiX image was 7.8 x 8.7 mm for each block. For the majority of blocks, MosaiX images from 2 areas were collected and RMax, RMin and Bireflectance images were calculated. Altogether more than 3,600 large MosaiX images were collected for the 21 cokes, which is more than 1,440,000 elementary images and over 1,300 hours of microscope time alone. Each acquired set was reviewed for quality and re-imaged if the quality or position on the block was not suitable.
Two types of segmentation were applied during optical image analysis: structural and textural. Structural characterisation segmented unreacted IMDC (Inert Maceral Derived Components), partially reacted IMDC, RMDC (Reacted Maceral Derived Components) and porosity. Textural characterisation segmented isotropic inerts, fused inerts, and fused vitrinite with high, medium and low bireflectance.
The optical image analysis revealed many structural features which were significantly different for the sets of matched cokes. It should be noted that the amounts of cokes analysed did not allow making of significant conclusions about Northern Hemisphere cokes and Australian cokes in general, so any references to some significant differences between them throughout the report are only applicable to the analysed sets.
Some bias was introduced by attempting to best match medium/high rank cokes, which only was revealed in the final stage when all imaging, image processing and calculations were performed
It should be noted that for all 15 textural parameters where 95% confidence interval for Northern Hemisphere cokes was different from Australian cokes, and at the same time significant (95%) correlation with RMax was observed, the difference between Northern Hemisphere cokes and Australian cokes was opposite to what would be expected if higher RMax in Northern Hemisphere cokes affected the result; such behaviour would be seen if the actual RMax for Northern Hemisphere cokes was lower than that for Australian. Such an outcome can possibly be the consequence of the fact that Northern Hemisphere cokes have different dependences between structural parameter and coal characteristics compared to Australian cokes.
The performed analysis for cokes made from medium/high rank coals demonstrated that porosity within IMDC in the tested group of Northern Hemisphere cokes is lower than in the Australian set of matching cokes.
It also was demonstrated that wall thickness (taking nodes into account) in Northern Hemisphere medium/high rank cokes was higher than in Australian cokes due to the fact that fine porosity in Northern Hemisphere cokes was lower. If nodes are not considered and fine porosity removed, the modified wall thickness (nodes removed) and neck thickness are not significantly different. Although, specific neck thickness (per unit area) is at 90% confidence higher for Australian cokes, meaning more connections between nodes, stronger coke.
There was also some indication that the average size of IMDC in Northern Hemisphere cokes (medium/high rank) was lower than in Australian cokes; they were less elongated and the boundary of IMDC in Northern Hemisphere cokes was smoother than in Australian cokes. The fact that the average thickness of voids between walls of RMDC connected to IMDC grains was larger for Northern Hemisphere cokes (medium/high rank), may mean that in Australian cokes RMDC adheres better to IMDC and so Australian cokes are less prone to cracking.
The authors tended to connect the amount of partially reacted IMDC with the capability of inerts to fuse. Calculations showed that at 90% confidence level, Northern Hemisphere cokes (medium/high rank) have more partially reacted IMDC than Australian cokes, which contradicts with the findings of Diessel and Wolff- Fischer (1987). They declared that some inertinites in Australian coals (medium/high rank) are more fusible than the inertinites in carboniferous coals from the Northern Hemisphere. This observation needs special understanding and analysis of how partially reacted IMDC are segmented during image processing and how they are connected with fusible IMDC in coke.
Analysis of porosity for cokes made from medium/high rank coals showed that porosity in Australian cokes is more tortuous than in Northern Hemisphere cokes and/or pore roundness is higher for Northern Hemisphere cokes.
Weak area analysis showed that Northern Hemisphere cokes (medium/high rank) had a lower average area of Weak areas and more rounded pores than Australian cokes after the Weak area procedure was performed, which may mean that Northern Hemisphere cokes are less prone to cracking from the point of view of porosity structure.
Separated pore analysis suggested that individual pore pockets in Northern Hemisphere cokes (medium/high rank) are larger than in Australian cokes which may have a negative effect on Northern Hemisphere cokes strength.
To summarise the above findings, image analysis showed that cokes in Australian set (medium/high rank) have finer porosity with larger tortuosity, but also more connections between nodes and better attachment of RMDC to IMDC than in Northern Hemisphere set.
This research clearly demonstrated that the structural and textural methodology for coke characterisation developed by CSIRO can comprehensively describe the coke texture and structure and can be successfully applied when comparison of textures and structures of different cokes/materials is needed.