STATISTICAL-MODELS FOR BITUMINOUS COAL - A 3-DIMENSIONAL EVALUATION OF STRUCTURAL AND PHYSICAL-PROPERTIES BASED ON COMPUTER-GENERATED STRUCTURES

Any coal structural model is at best an average representation. Even within a single maceral and rank there are numerous possible models. We have recently developed a computer technique which determines the number of different models which can be constructed from a set of analytical structural data. In the present work, the technique is applied to vitrinite from high-volatile C (HvC) bituminous coal. The number of possible vitrinite models is too large to build all of them and too large to select one as representative of all the possibilities. However, we use sampling theory to define an optimal set of models (a sample) which is an accurate representation of the entire population of possible models. Physical characteristics such as the potential energy, the helium density, and the microporosity are then calculated for the sample of vitrinite models, extrapolated to the whole population, and compared with the corresponding experimentally-determined data.