CSR Predictions and Correlations Definition

Coke Strength after Reaction (CSR) is arguable the most price specific indicator of coking coal, but is difficult to test from samples such as borecores due to its mass requirements and high cost in testing.

Prediction of CSR from more routine coal quality analysis has been undertaken in several published studies, outcomes usually correlated to multiple coal quality parameters, including deterioration prone properties that are not readily applicable to Borecore outcomes. The published equations derived and used for the source coal rarely had accuracy defined and have been found by MCQR when applied to different deposits to have unacceptably large bias values in particular, and precision values that are not indicative of real outcomes.

This project's objective was to statistically evaluate CSR predictor performance from various coal quality CSR proxies (CSR Predictors) and define methods for future deposit evaluations. Data sets for multiple deposits were attained containing borecore ply / seam based information and accompanying production / despatch (shipping) data. Further, data for coking products on a deposit basis from project C17053 was also attained and evaluated. Published equations for CSR Prediction were also compared.

Univariate correlations for each dataset were first established and correlation coefficient (R Square or R²) / standard error outcomes observed for the coal quality tests examined (nil bias exhibited as data was fitted). Multivariate inputs and equations were then derived based on the univariate outcomes and statistically compared for bias and variability. Several different coal quality combinations were considered, and guides for new deposit evaluation established from the processing undertaken, including coal quality selection proxy for application (borecore or production / shipping) - samples with deterioration prone properties being highlighted for use only in immediate testing environments (production and shipping).

The outcomes of the project were the definition of accuracy information for varying CSR predictors with differing coal types, and method rules for creating new CSR predictors for new deposits. This aids industry in correct resource representation and / or lower analysis costs and better quality control. The outcomes then providing a financial risk reduction and / or evaluation enhancement.

This report forms a guide for use in industry to create and define CSR predictor methods and accuracy and thus resource / reserve / marketing product potential and confidence.