A mixed integer model for optimizing equipment scheduling and overburden transport in a surface coal mining operation

Abstract

Recently, competition has increased in the surface coal mining industry, which has necessitated the development of more efficient methods for coal removal. Despite this trend, very little emphasis has been placed on the development of optimization techniques applicable to the surface coal industry. The available methods are inadequate in that they recognize neither the complex equipment interactions present in a surface mining operation nor the interdependence of overburden removal and spoil placement.

The lack of available techniques prompted the development of a mixed integer model to optimize the scheduling of equipment and the distribution of overburden in a typical mountaintop removal operation. Using this format, a (0-1) integer model and transportation model were constructed to determine the optimal equipment schedule and optimal overburden distribution, respectively. To solve this mixed integer program, the model was partitioned into its binary and real-valued components. Each problem was successively solved and their values added to form estimates of the value of the mixed integer program. Optimal convergence was indicated when the difference between two successive estimates satisfied some pre-specified accuracy value.

The performance of the mixed integer model was tested against actual field data to determine its practical applications. To provide the necessary input information, production data was obtained from a single seam, mountaintop removal operation located in the Appalachian coalfield. As a means of analyzing the resultant equipment schedule, the total idle time was calculated for each machine type and each lift location. Also, the final overburden assignments were analyzed by determining the distribution of spoil material for various overburden removal productivities.

Subsequent validation of the mixed integer model was conducted in two distinct areas. The first dealt with changes in algorithmic data and their effects on the optimality of the model. The second area concerned variations in problem structure, specifically those dealing with changes in problem size and other user-inputted values, such as equipment productivities or required reclamation. For each of these optimal schedules and assignments obtained from the model, analyses were conducted in manner similar to that discussed above.