INVESTIGATING RELATIONSHIPS BETWEEN METHANE EMISSIONS AND ATMOSPHERIC DATA IN UNDERGROUND COAL MINES TO DEVELOP A FORECASTING MODEL

Big data is generated from both surface and underground mining operations. Such data contain a wealth of information concerning safety and health in the workplace and production parameters. This paper will discuss the progress towards developing an accurate forecasting model for methane gas concentration based on the analysis of data collected from Atmospheric Monitoring Systems in underground coal mines employing time series models. Several procedures need to be applied to raw data, such as data cleaning and filtering for outlier removal, data homogenization and segmentation, and exploratory statistical analysis. The data analyzed were retrieved from two coal operations in the Eastern US. A negative correlation between barometric pressure and methane gas concentration was established, at least for certain data segments. Such correlations raise the possibility that barometric pressure data can predict variations of gas concentration in the mine. The datasets will be further investigated to establish the robustness of barometric pressure and gas concentration correlation and explore the dependence of gas concentration on other factors related to mine design and operations.