MULSIMNL/LARGE: REVIVING A USBM TOOL FOR MODELING COAL MINES

摘要

Specialized numerical modeling codes have been developed over the years to estimate ground stresses during the excavation of coal in underground mines. One of these tools, MULSIM, a displacement discontinuity boundary element code, was last developed by Zipf into a version called MULSIM/NL [USBM 1992a, 1992b]. The code was used frequently in years past, but closure of the Bureau of Mines in 1996 marked the beginning of a pause in its development. A similar tool with distinctly different treatment of overburden idealization, LaModel, was subsequently developed [Heasley 1997] and largely displaced MULSIM/NL in the literature. However, recent work modeling deep western longwall coal mines [Larson and Whyatt 2013] has demonstrated that MULSIM has distinct advantages for regions with relatively stiff and strong overburden. This work required that MULSIM be revived and updated into a version called MulsimNL/Large. Updates include larger array sizes, greater precision, greater number of materials, and the addition of a five-point, six-segment pillar load/displacement model that provides the user a more flexible coal-strength model allowing hardening and softening before a final residual strength. MulsimNL/Large has the ability to calibrate to large load transfer distances observed in some western U.S. coal mines that appear to be caused by strong, massive units in the overburden and/or floor of a coal mine. Also, it provides the ability to calibrate gob stiffness to simulate measurements or estimates of overpanel weight distribution between gob and abutment, while still maintaining load transfer distance with minor adjustments. MulsimNL/Large can easily implement material strengths calculated from various pillar strength models. The five-point model gives MulsimNL/Large the ability to easily simulate pillar behavior that is extracted from detailed pillar models, such as FLAC3D. MulsimNL/Large or similar approaches should be considered for evaluating mining layouts where stratigraphic conditions are known to cause overpanel weight to transfer to an extent significantly longer than the empirical method widely used to calculate load transfer distances in the U.S. [Peng and Chiang 1984; Mark 1987; USBM 1990].