Analysis of incidents in underground mining reveals that ground falls, including roof and rib falls, is one of the major sources of injuries and fatalities, however the number of accidents has reduced significantly during recent years (MSHA 2015). For instance, in 2005, around 60% of total fatalities in underground mining were associated with ground falls. This percentage has dropped to 37% in 2015 (MSHA 2015). This significant reduction in underground incidents may be caused by several reasons such as: using advanced technology, advance in computer science, new policies on safety, etc. Development of mathematical methods such as finite element and boundary element methods can be classified under new advanced in computer science provided new computer programs for using in the mining industry. In general, these programs are used as a powerful tools to enhance the safety of mine designs. There are several programs have been developed and each program provides specific features and strengths as well as some specific limitations. The LaModel program was developed based on the displacement discontinuity method and has been successfully used over several years in mining industry.;During the last few years, several features have been added to the LaModel program to increase the ability of the program to analyze the underground coal mines. The new efforts lead to implementation of several new options into the program such as: gob modulus calibration, a local mine stiffness calculation for evaluating the potential of bump, and adding a subsidence calibration technique, and the energy release rate calculation as well as calibration of the model These components increase the analysis ability of the program for several specific conditions.;Enhancing the main algorithm of solution for mines with multiple seams is necessary to help mine design easily. Presently in the LaModel program, the fundamental elliptical, partial-differential equation of the laminated overburden is solved using a finite-difference approach. This finite-difference solution has proven to be fast when solving single seam problems. However, when solving multiple-seam problems, the solution time increases by several orders of magnitude. The multiple-seams model in LaModel are solved by first calculating the seam convergence distribution seam by seam and then calculating the multiple-seam stress between the seams. Calculating and projecting stresses between the seams is the most time consuming process in a Multiple-seam model. In this dissertation, three algorithms have been developed and implemented to improve the efficiency of the multiple-seam calculation in the program. These algorithms include: a new kernel integral formulation, an optimized influence distance algorithm and a double macro element method. These new algorithms have been implemented into the most recent version of LaModel. Initially the algorithms were validated by comparing old model results against new model results. Then, the speed of the new algorithms were benchmarked, both individually and in combination, using 5 different case histories. Finally, the combination of these three algorithms has been implemented into the program and accuracy of the final version has been evaluated. By implementing and using new method, the LaModel program will be more efficient than current version of program in terms of multiple seam computations. It means the model with more details and large size can be solved in less time in comparison with current multiple seam solution. Therefore, the major contribution of the current research in ground control field is improving safety in underground mining in multiple seam reserves.
展开▼
