Using Geostatistics To Predict The Characteristics Of Washed Coal - Introduction

The increasing cost of developing new mines and the need to mine more marginal deposits have led mining companies to use more sophisticated estimation methods such as kriging. But until recently, coal mining companies have shown little or no interest in these techniques. However, the stringent quality controls which are now placed on coal have led to a surge in interest in improved estimation methods. This can be seen from the -, number of research projects by private companies and governmental agencies on the application of geostatistics to coal reserve estimation. At present, most of these programs are centered on the problem of estimating the tonnage and the average grade (i.e. ash percent, sulphur content, calorific value, etc.) of the coal in-situ. See, for example, Sabourin (1975), Woods (1976, 1979), Pauncz and Nixon (1979) and Armstrong (1980). These studies show that the established techniques of linear geostatistics can be applied successfully to coal deposits. The study by Woods is particularly interesting because he had access to the actual production figures as well as the initial sample data from the no 2 seam in the Greenside Colliery owned by the Anglo-American Corporation. He showed that the estimates obtained by kriging were consistently closer to the actual production figures than those given by three other estimation methods. This confirms the superiority of kriging as a method of estimating block values for coal. Another advantage of geostatistics is that it can be used to determine the optimal spacing between drill holes. An example of this is given by Armstrong for the coking coal deposit in the Bowen Basin, Australia.