Estimating the probability of satisfactory conditions for main entry intersections in a coal mine in weak rock under deep mining depth

Abstract

The satisfactory condition of underground coal mine intersections is a fundamental requirement for the safe operation of any underground coal mine. A satisfactory intersection condition goes beyond prevention of a roof fall. It must include the ability of the intersection to safely perform its designed function. A satisfactory condition is especially important for those main entries that provide life-of-mine access; both to the current, active areas of the coal mine, and to the future mining areas of the coal reserve or resource. Life-of-the-mine access may be required for decades. The intersections in the main entries must provide safe, satisfactory conditions and performance during this time. Long-term, satisfactory intersection conditions are particularly problematic in a deep/weak rock coal mine environment characterized by: (1) mining depths greater than 800 ft, (2) a poor to fair quality rock mass above the coal seam (RMR ≤ 50); and, (3) very weak to weak rock strength conditions for the immediate rock strata above and below the coal seam (CMRR ≤ 35). This research collected, compiled, and analyzed a data set of the geometries, environments, and conditions of 884 intersections in the main entries of a longwall coal mine, operating in the deep/weak rock coal mine environment. The database compiled from the data set included binary, categorical, and numeric predictor variables. The research developed a binary outcome variable from the database, an intersection condition rating as Satisfactory or Not satisfactory, that included the design function requirements of each intersection over the life of the intersection. The data were analyzed and a list was compiled of statistically significant predictor variables impacting the satisfactory condition of an intersection at the mine. An intersection condition probability estimation model was developed from the derived outcome variable, the list of predictor variables, and the data set. The probability estimation model was derived using logistic regression. This research was unique because it collected a large database specific to the deep/weak rock coal mine environment in the main entries of an operating, longwall coal mine in the US. It is the first application of the logistic regression method to derive an equation for estimating the probability of satisfactory intersection conditions. The research developed a novel method for the in-mine evaluation and rating of the existing safety and conditions of a coal mine intersections that incorporates the conditions of the roof, ribs, floor, secondary and supplemental support, and function performance for use by the coal mining industry.