Coal mine gas originates from the gas emission zone (GEZ) of the mine, as well as the longwall face and pillars. Gas emissions are controlled directly at the sources using horizontal or vertical boreholes drilled from surface or from the entries in advance of mining, or it is captured from the fractured and caved zones (gob) using ventholes during mining. The rest of the gas, especially that gas that originates from the longwall face and caved zone, mixes with the ventilation air and travels through bleeder and return entries before being exhausted to atmosphere from air shafts.

Although the gas associated with mining mostly focuses on methane, the gas is not pure methane but is a mixture (where both the components and their quantities vary depending on the sampling location). Understanding the evolution of the composition of the gas from source to different sampling and evaluation points in mines using proper statistical analysis and interpretation methods can lead to better designed degasification and ventilation systems and the selection of the most adequate utilization method for generating energy from the gas.

In this work, we present the results of compositional data analysis (CoDa) of gases sampled at different locations in a longwall mine operating in Pennsylvanian coal-bearing strata in the Northern Appalachian coal basin. Sampling locations were from accessible parts of the bleeder entries, returns, bleeder evaluation points and shafts within the mine, and also from gob gas ventholes (GGV) and coal degasification boreholes drilled in the panel areas. In addition, desorbed gas samples from seams that are important for the mine were included in the analyses for comparison. The compositional data analysis showed that the gas composition shifts based on the sampling location, putting in-mine and pure coalbed gases on the opposite ends. Removal of air from the samples did not change this observation suggesting that oxygen is depleted especially for in-mine samples, which is due to oxidation. Results also suggested that desorbed gas samples may not represent the composition of the coal seam gas.