ACARP Project Number: C20009      Published: May 13

 Philip Bennett, Frank Shi, Nick Andriopoulos

Coke quality is commonly characterised by tumble drum strength tests, typically on coke produced in pilot scale coke ovens. The empirical nature of the commonly used suite of tests usually requires that all the major coke tests need to be completed for a single coke. Additionally, the variability of each test measurement can be high, and comparison between different tests remains problematic. It is becoming more common for coking coal producers, cokemakers and ironmakers to try to understand how different blending or washing practices will influence coke quality. The identification of increasingly subtle potential improvements or deteriorations in coke quality is clouded by the deficiencies in drum tests.

This problem is further exacerbated by different sub samples of the same coke being used in the different drum tests and differences in the feed sizing. It is reasoned that the difference in the drum indices obtained by various drum tests is caused by the difference in applied tumbling energy due to variations in drum diameter and tumbling time.

This project proposed that a unified index or indices can be generated once the size reduction and energy relationship is established and the applied energy in various drum tests is corrected.

 The project investigated whether the coke size data, normally collected when a set of drum tests are performed, can be used to determine a theoretically based strength index based on sound breakage theory and, since it is determined from all the data generated, is a better representation of the coke strength for the whole coke sample than is possible with individual drum indices.

The objectives of this project were to:

The JKMRC developed a unified Coke Strength Index (CSI) from a theoretically based breakage model which takes into account the coke feed size distribution, product size distribution and the total effective energy applied in a range of drum tests. It has been demonstrated that this model agrees with the work of other researchers for volume breakage and can be used to predict the different volume breakage drum indices (such as Stability, M40, I40, I20, DI3015and DI15015) used by the international steel and coal industries. This project has demonstrated that the coke size data, normally collected when a set of drum tests are performed, can be used to determine the CSI and, since it is determined from all the data generated, it is a better representation of the coke strength for the whole sample than individual drum indices.

The use of the CSI will greatly assist in the identification of any potential improvements or deteriorations in coke quality where the coke strength data is clouded by the deficiencies in drum tests. The results of this work can make it possible to develop a drum testing procedure with smaller amounts of coke being required and still being able to calculate all the volume breakage drum indices. This can be done using smaller compartments within pilot scale coke ovens, for example using a multi-compartment coke charge. While the amount of coke produced does depend on the yield of coke and mean coke size after stabilisation, a revised coke testing procedure will reduce the coke requirements needed to determine the tumble drum strength of cokes and therefore lower the cost of the coking coal evaluation.