Comparison of Underground Coal and Trona Mine Seismicity

"BACKGROUND Strata mechanics of three underground coal mines in central Utah were recently studied using a specialized finite element technique [1, 2, 3, 4]. The purpose of these studies was to examine relationships between mining and related seismicity. The longwall mining method was predominant, but room and pillar mining and barrier pillar mining were also used in the mines studied. In every case, high correlation between face advance and seismicity was observed. The specialized finite element technique allows for whole mine subsidence estimation at a kilometer scale and for stress concentration at a meter scale near a working face. Comparisons of observed mine subsidence with model subsidence allows for model verification and justifies detailed analysis of the evolution of stress at seam level. Yielding is expected near a longwall panel face and as the face is advanced the extent of yielding is expected to increase. The total number of element failures in a finite element mine model also increases, so a high correlation with associated seismicity, say, event count, is not so surprising. A brief review of recent coal mine studies in Utah is presented in the sequel. The procedures are given in some detail in application to trona mining. COAL MINES The analysis procedure begins with the geologic column and proceeds to specification of strata properties including elastic moduli and strengths. Joints are an integral part of the geologic model and joint set specifications include dip direction, dip, spacing, stiffness and strength. Spatial variability of strata properties in situ may be taken into account during analysis. Topography is important, especially to Utah coal mines that are developed from outcrop in regions of high relief. Mesh generation follows specification of data input. Finite element analysis is next, followed by reporting of results. In this regard, all finite element computations are based on first principles. No model “calibration” is done that involves adjusting computer input to obtain a fit to mine measurements such as subsidence profiles. Figure 1 illustrates the concept of a dual node – dual mesh finite element concept. Dual nodes are roof and floor nodes in mined areas. These nodes replace the seam. The dual mesh is a much refined mesh with displacement boundary conditions obtained from the big mesh and thus replaces the effect of mining in the region outside the dual mesh. The refined dual mesh allows for details of stress, strain and displacement near a working face of interest. Strata Properties Strata properties include elastic moduli and strengths of intact rock between joints and joint stiffness and strength for each joint set present. Both properties sets may be obtained from laboratory test data, although testing joints poses a sampling problem in the field. The statistical variability observed in laboratory test data, say, in unconfined compressive strength tests is often characterized by much variation. In fact, the coefficient of variation, ratio of standard deviation to mean, is often greater than 35 percent. When sample values are mapped back to the mine location, one may observe spatial variability as properties vary from place to place as shown in Figure 2. Spatial variability may be incorporated into a finite element analysis with probabilistic assignment of properties to elements in accordance with measured variability. The result is a more realistic model of rock in situ."