Factors Affecting Coke Size and Fissuring during Cokemaking

The literature on the mechanism of fissuring during coking was reviewed and alternative approaches to the modelling of the fissuring mechanism evaluated. A new mechanism and model for the physics of fissuring was proposed. The model incorporates:

• The primary fissure pattern is determined early in the coking process
• Some fissures extend and some cease propagating (or propagate more slowly) as coking continues. The fissure spacing and penetration depth can be estimated using an energy minimisation technique.
• The modelling and experiments indicate that the optimum fissure spacing increases with time through the coking process. When the optimum fissure spacing increases sufficiently alternate fissures stop propagating to accommodate the larger required spacing.

The model qualitatively explains a number of features observed in experiments where the coke oven was pushed soon after charging the oven. The techniques developed for pushing the coke oven only a few hours after charging have the potential to allow further studies on the structure of the semi-coke - plastic layer interface to improve understanding of the formation of coke structure, strength and size. The model requires further work to improve the quantification of the analysis and to better define the criteria for when fissures cease propagating.

The model identifies key coal factors affecting fissuring as strength of the semicoke and the temperature related shrinkage characteristics of the coal. Further experimental work is required to better characterise each of these.

Image analysis techniques were modified to allow the characterisation of the fissure patterns of intact coke cakes pushed from the pilot coke oven. The relations between fissure spacing, coal shrinkage and coke size were studied. Relations were seen between fissure spacing, shrinkage and mean coke size (or xbar in the Rosin Rammler distribution). The distribution of the coke lump sizes was more difficult to model. A weak relation was observed between the Rosin Rammler slope and shrinkage. There was not a clear relation between Rosin Rammler slope and fissure spacing or distribution of fissure spacings. Improved measurements of shrinkage over the Sapozhnikov method used in this study are required to fully define the shrinkage curves required for the model. High temperature dilatometers provide a possible method for obtaining the required information.

Statistical analysis of pilot coke oven data was performed to identify coal factors that affect mean coke size and the coke lump size distribution, represented by the Rosin Rammler distribution. The experiments were performed under constant coking conditions and with a uniform bulk density distribution to minimise the effect of operational factors on the analysis.

The mean size of the coke and the xbar for the Rosin Rammler distribution of lump coke were primarily determined by the blend rank and one of several possible measures of plastic properties of the coal. The Rosin Rammler slope parameter was more difficult to model. Important factors affecting the slope were the plastic properties of the coal (mainly fluidity) and vitrinite. More work is required to understand the relationship between the Rosin Rammler slope and coal properties.