Panel-scale modeling of a deep longwall panel: the MULSIM alternative.
Proceedings of the 32nd International Conference on Ground Control in Mining, July 30-August 1, 2013, Morgantown, West Virginia. Morgantown, WV: West Virginia University, 2013 Jul; :181-188
Panel-scale modeling of stress changes induced by longwall coal mining is a challenging problem, particularly in deep mines of the western United States. Two common approaches to this problem are the empirical and boundary element modeling methods. One boundary element method, LaModel, was recently calibrated to field measurements at a mine in Colorado (Larson and Whyatt, 2012). However, the calibrated overburden properties were extreme, resulting in very low estimates of gob loading. Next, the empirical method was calibrated to field measurements and applied to the problem, but this approach clearly extrapolated this method well beyond the characteristics of underlying cases. This paper examines a third option, the elastic continuum boundary element program Mulsim. The program was expanded to handle LaModel-size meshes in a version dubbed Mulsim/ Large. Properties for the elastic, homogeneous overburden model in Mulsim were estimated as a thickness-weighted average of typical published elastic properties for each major stratum. These properties, along with a commonly used pillar strength equation and coal properties, produced the measured load transfer distance and reasonable gob loading without further calibration. Given the number of input parameters required for all of these methods, it is not surprising that LaModel and the empirical approach could be adjusted further to incorporate similar gob loads. However, these extraordinary measures failed to bring gate road pillar loading in line with Mulsim results. Differences in pillar load estimates continued to be significant (as large as 65% in this study). These results demonstrate that, for this particular case, the choice of analysis method impacts results in ways that cannot be removed by calibration of input parameters. Finally, the natural fit between the Mulsim model and mine conditions suggests that approximating overburden at the particular site used for this research as an elastic continuum is superior to the alternatives examined.
Mining-industry; Underground-mining; Ground-control; Ground-stability; Geology; Computer-software; Computer-models; Models; Coal-mining; Longwall-mining
Proceedings of the 32nd International Conference on Ground Control in Mining, July 30-August 1, 2013, Morgantown, West Virginia