This is a past event.

Modeling compositional data for characterizing geologic resources: An example from mapping coal ultimate analyses data

The compositional properties of coal can significantly influence how it should be mined, handled, processed, utilized, and even how coal-fired power plants should be designed to reduce emissions. Ultimate analysis quantifies principal chemical elements in coal as parts of a composition. These elements are C, H, N, S, O, and, depending on the basis, ash, and/or moisture. The treatment of parts using proper compositional methods may be important for mapping them free of mathematical artefacts, as most mapping methods carry uncertainty due to partial sampling, which increases spatial fluctuations, and thus errors.

In this work, we mapped the ultimate analyses parts of the Springfield coal from an Indiana section of the Illinois basin, USA, using sequential Gaussian simulation of isometric log-ratio transformed compositions. We compared the results with those of direct simulations of non-transformed compositional parts. We also explored the implications of these approaches in mapping other properties that rely on compositions through correlations to identify the differences and consequences.

This study demonstrated that by using compositional simulation and sequential Gaussian simulation, the mathematics and the limits of the data are not violated. In our study, compositional simulation gave exactly 100% as the sum of values of parts, as expected, regardless of the type of the map and how many realizations are included in assessment. Direct modeling of parts, using variograms of raw data, on the hand, generated as much as ~+/-21% error in sum of the part values in individual realizations.

Further, application of part values modeled using both approaches were also used in a correlation to predict calorific value of coal. The results were compared with modeling of collocated pointwise calorific value data. Results showed that the data spread was closer to the data range from calorific value maps predicted using parts from compositional modeling, which showed less prediction error compared to direct geostatistical modeling of parts.

This study shows the practical advantages of utilizing compositional data analysis in mapping of coal properties and in obtaining mathematically consistent results. The approach can also be extended for mapping saturation distributions, and fluid and rock composition distributions in compositional reservoir simulations of coalbed methane and enhanced recovery processes.