Quality management in a 3D geological model – reliability of predicted hard coal quality parameters

Abstract

This paper presents some of the results of the project undertaken in JSW SA, which aim was to create a three-dimensional model of the deposits that make up the company and schedule company’s production. The assessment of the quantity of coals without analysis of qualitative data, i.e. physicochemical parameters, coking parameters, and optical petrographic analysis is not suitable for obtaining commercial contractors. To obtain information on the quality of the coal seam, the geological service of the mine takes coal samples. In the stratigraphic model and quality model, dedicated interpolators are used for interpolation and extrapolation. In the seam quality model, the most optimized interpolators are Inverse and Height. Modelled parameters such as volatile parts content and random vitrinite reflectivity were analysed in detail. The Height interpolator looks for both a random and a linear dependency. It extracts random changes locally while searching for linear dependencies and extrapolates them to a deposit area that does not have qualitative data. There is a risk of extrapolating a given value to infinity. Nevertheless, the amount of data and the area modelled allows you to close its scope at an acceptable level. A separate POLYGON interpolator based on mxl express surfaces was created to map coal type range. It uses interpolated quality parameters at a given location, generating a range of a particular type of coal. Setting the trend of variability makes it possible to predict higher coal types in deeper, unrecognized batches of deposits according to documented variability in parameters.